Geospatial approaches for mapping zero-dose children in low- and lower-1 middle-income countries: A scoping review

Geospatial approaches for mapping zero-dose children in low- and lower-1 middle-income countries: A scoping review | 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 Systematic Review Geospatial approaches for mapping zero-dose children in low- and lower-1 middle-income countries: A scoping review Ann M. Njogu, Moses M. Musau, Swati Srivastava, Francesco Menegale, and 16 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9304718/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: Zero-dose (ZD) children remain a critical public health concern, particularly in low- and lower-middle-income countries (LLMICs), where over 80% of the global ZD population resides, disproportionately concentrated among the most marginalised. Geospatial methods have emerged as effective tools for identifying and targeting immunization gaps. However, no review has systematically documented the spatial data and methods used to identify and characterize ZD children and corresponding gaps. This scoping review addresses that gap across LLMICs. Methods: Following PRISMA-ScR guidelines, we searched for peer reviewed articles published upto 2025 on spatial modelling of childhood vaccination coverage in LLMICs using six databases: PubMed, Web of Science, Scopus, Cochrane, Embase, and EBSCOhost-CINAHL. We extracted details on study characteristics, covariate types and sources, modelling methods, and the gaps. Articles were thematically summarized focusing on geospatial data, modelling approaches, and their corresponding gaps. Results: Of 15,587 articles retrieved, 102 from 68 LLMICs were included, with 70% published between 2021 and 2024, and 87% concentrated in Ethiopia, Nigeria, and India. Only a fifth assessed ZD prevalence, based on two distinct definitions. Studies relied predominantly on household survey data, with routine administrative data underused. Covariate data were dominated by demographic factors (49%) with limited representation of hard-to-reach contexts. Methods included clustering and autocorrelation analysis (54%), spatial interpolation (45%), small-area estimation (13%), and machine learning (8%). Key gaps included inconsistent ZD definitions, missing covariates, data inaccuracies, sparse samples, and weak representation of conflict-affected, informal-settlement, and mobile populations. Conclusions: Despite growing availability of spatial data and methods, geospatial identification of ZD children remains concentrated in few countries, relies heavily on survey data, uses inconsistent definitions, and is constrained by limitations that systematically exclude the most marginalised populations. Addressing these gaps will require harmonised definitions, integrated data systems, and reproducible modelling approaches underpinned by sustained investment in local analytical capacity. Epidemiology Statistical Epidemiology Biostatistics Medical Informatics Health Policy Geographic Information Systems zero-dose immunization coverage vaccination inequity spatial high-resolution geostatistics geospatial modelling mapping small-area estimation data Low- and lower-middle-income countries Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Childhood immunization is a pillar of public health, significantly reducing morbidity and mortality from vaccine-preventable diseases (VPDs) (1). Since the establishment of the World Health Organization‘s (WHO) Expanded Programme on Immunization (EPI) in the 1970s, there has been enormous progress in expanding access to life-saving vaccines worldwide, with 85% of children worldwide now receiving the third dose of diphtheria, tetanus, and pertussis (DTP3) containing vaccine, a common measure of immunization system performance (2). However, even in countries with high immunization coverage levels, there are still pockets of zero-dose (ZD) children - those who do not any receive routine immunization (3,4). Operationally, ZD children are defined as those missing the first dose of DTP (DTP1) vaccine by the end of their first year of life (4). This definition is used because DTP1 is universally used in routine immunization, recommended at six weeks, and is thus a reliable indicator of initial contact with routine immunization services (5). ZD children are consistently under-served across multiple dimensions of the health system. The absence of DTP1 is not only an indicator of missed immunization but also serves as a practical proxy for limited or absent contact with routine health services more broadly (4). As such, ZD status often reflects wider patterns of exclusion, including reduced access to primary healthcare, maternal and child health services, and other essential interventions. The multiple challenges that ZD children face, for example, nutrition, education, water and sanitation, and other essential resources, exacerbate their vulnerability and risk to disease(6). ZD children remain vulnerable to VPDs, which can contribute to sustained disease transmission, even in populations with high vaccination coverage (7). By 2021, approximately twenty-five million children worldwide were not fully vaccinated, with 73% classified as ZD (8). Over 80% of these ZD children were in low- and lower-middle-income countries (LLMICs), with sub-Saharan Africa (SSA) bearing a disproportionate burden. India, Nigeria, Ethiopia, the Philippines and the Democratic Republic of Congo (DRC) accounted for 15%, 12%, 6%, 6%, and 3.8% of the global ZD population, respectively (9,10). Large pockets of ZD children in LLMICs are located especially among marginalized communities (11). For example, in 2024, 31% of children under five years in Nigeria were ZD, clustered in the northwestern parts especially in Zamfara, Sokoto and Kebbi states, and in the poorer and less educated subgroups (12). Years of suboptimal vaccine coverage have resulted in an accumulation of under-immunized populations, fuelling outbreaks of VPDs such as measles, whose global cases reached 9.8 million in 2022, leading to about 136,000 deaths (13). The COVID-19 pandemic accelerated global vaccination challenges, leading to declines in immunization rates. In India, for instance, the coverage of DTP3 dropped from 91% to 85% during the COVID-19 pandemic (14). This reversal temporarily erased over ten years of progress in vaccination, primarily due to pandemic-related disruptions such as lockdowns and the redirection of healthcare resources towards combating COVID-19 (15). Although the 10.9% rise in ZD children between 2019 and 2024 (16) is partially attributed to COVID-19 disruptions, other factors like poverty, inadequate funding for immunization programs, geographic inaccessibility of immunizing health facilities, armed conflicts, weak health infrastructure, vaccine hesitancy, and cultural or religious opposition to vaccination have been cited as long-standing barriers to reaching ZD populations in LLMICs (17,18). Consequently, global initiatives such as the Immunization Agenda 2030 (IA2030) (19) and the GAVI Strategy 5.0 aim to achieve immunization equity, with a particular focus on reaching ZD communities (20,21). The IA2030 Impact Goal Indicator 2.1 aims at reducing the number of ZD children by 50% by 2030 using 2019 as the baseline year, and targets to maintain a DTP1 coverage of 99% in countries where DTP1 coverage is already at that level (22). To achieve United Nations’ Sustainable Development Goal (SDG) 3: good health and well-being (23), it is therefore crucial to identify, target, and reach ZD children with equitable and cost-effective interventions to efficiently address immunization disparities (24–26). Gavi’s Identify-Reach-Monitor-Measure-Advocate (IRMMA) framework emphasizes that the first step in reaching ZD children is accurately identifying and characterizing the communities to which they belong (27). To identify and characterize these communities at high spatial resolution, spatial, statistical, or a combination of both approaches, have been employed (9,11,26). These approaches often integrate covariates (factors, determinants or variables associated with ZD prevalence or low immunization coverage) with vaccination coverage, to generate high-resolution estimates of ZD prevalence. These spatial and spatial-statistical techniques emerged as more effective tools for unmasking disparities in ZD prevalence that are often obscured by regional or national estimates (28,29). In the last two decades, spatial and spatial-statistical approaches used for mapping health outcomes have considerably advanced (30), including for vaccination coverage (31–33). This has been fuelled by the increasing availability of immunization data from routine health information systems (e.g., District Health Information Software-2 (DHIS2) and household surveys (e.g., Demographic and Health Surveys (DHS) and Multiple Indicator Cluster Surveys (MICS)) that incorporate global positioning systems (GPS) (34,35), as well as geocoded databases of health facilities (36), satellite derived variables and development of advanced methods to synthesize these datasets (30). For example, the number of countries using DHIS2 for health reporting increased from 11 to 99 between 2010 and 2025, while DHS surveys have increased exponentially between 1985 and 2025, now covering 92 out of 129 LMICs (37–39). However, there is a conspicuous gap in systematic synthesis of spatial data sources, covariates and modelling frameworks used to identify and map ZD and under-immunized children. Previous literature reviews have mainly focused on factors or determinants (40–42), barriers (43,44), and inequalities (45–47) associated with general childhood vaccination coverage or ZD prevalence. While findings of these reviews highlight disparities in vaccine coverage among vulnerable groups (such as migrants, refugees, and those in informal settlements), they do not have a geospatial lens, leaving a critical gap in high-resolution mapping of the heterogeneity in observed coverage levels and in understanding key spatial determinants. So far, only one narrative review has applied a geospatial lens to examine microplanning, geospatial, and machine learning (ML) approaches for reaching ZD and under-immunised children. While it provides a narrative overview of the strengths and limitations of these approaches, it was restricted to 18 publications and reports in SSA only mainly published in in 2024 and 2025, did not explicitly focus on spatial methods, and did not critically assess the methods, data and data sources used in the studies (48). Adopting a geospatial lens is critical for identifying ZD children in LLMICs, as they are often found in clusters that are masked by national or regional averages (49). Geospatial approaches enable the detection of fine-scale geographic inequities, revealing underserved communities affected by remoteness, conflict, urban informality, or weak health system reach (49,50). By integrating spatial analysis with demographic and health data, policymakers and practitioners can more precisely target resources, design context-specific interventions, and improve the efficiency and equity of immunization strategies aimed at reaching ZD children. To date, no synthesized evidence has described (i) spatial datasets that have been used to represent determinants of ZD and under-immunization, (ii) spatial and spatial-statistical approaches that have been applied to generate ZD estimates in LLMICs and (iii) gaps limiting policy-relevant use of these approaches. To address these gaps, we conducted a scoping literature review covering all LLMICs and outlined the relevant spatial datasets that have been used to map ZD prevalence. Secondly, we highlighted the methods used in mapping ZD children. Further, we outlined gaps in both geospatial data and methodological frameworks used, and the corresponding recommended strategies to address these gaps in most vulnerable populations. A scoping review is appropriate because evidence on geospatial methods for identifying ZD children is diverse, cross-disciplinary, and evolving. Research varies in terms of data sources, characteristics, and analytical approaches, making traditional systematic reviews impractical. A scoping review enables mapping of the literature, identifying key concepts, gaps, methodological trends, and how geospatial approaches are applied to study ZD populations. Materials and Methods Search Strategy We adhered to Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews (PRISMA-ScR) guidelines (51) and the protocol was registered in Open Science Framework (52). A comprehensive search for peer reviewed articles was conducted in April 2025 in six electronic databases: PubMed, Web of Science, Scopus, Cochrane, Embase, and EBSCOhost-CINAHL. Search terms were designed in accordance with the Population-Context-Concept (PCC) framework (53), where Population encompassed all studies about vaccination coverage; Concept referred to studies using spatial or spatial-statistical modelling methods; and Context to studies in LLMICs (Supplementary file 1 table S1). Screening and selection criteria Article screening was conducted in two phases: title and abstract screening and full-text review. Screening was conducted independently by two reviewers (AMN and MMM), to minimize bias in the selection of articles. Any disagreements regarding exclusion or inclusion of articles were resolved by a third reviewer (PMM). We included studies published in English that were conducted in LLMICs and used, or explicitly described the use of, spatial data to map or model ZD children or routine childhood immunization coverage with no temporal limits. We also included studies using geospatial methods to predict, visualize, or analyze coverage of routine childhood vaccinations such as DTP1, measles-containing vaccines (MCV), among others. Studies were excluded if they were conducted outside LLMICs, focused on immunization coverage without specific attention to childhood vaccination, or addressed only clinical aspects of vaccine efficacy or safety without examining coverage. We also excluded literature reviews and studies that examined determinants, barriers, or correlates of immunization uptake without incorporating a spatial component. The inclusion and exclusion criteria are detailed in Supplementary file 1 table S2. Although the focus of this study was primarily on ZD children, studies assessing child-routine vaccination coverage through geospatial techniques were also retained since the same data and methods apply. Search terms were first piloted on PubMed (Supplementary Box 1) and then adapted to the other databases. Additional grey literature was retrieved from unpublished reports, conference proceedings, websites, dissertations and snowballing of eligible studies(53). Data Extraction From each study, we extracted: i) bibliographic information, including title, authors, Digital Object Identifier (DOI), year of publication, and study objective, ii) study characteristics such as the objective, geographical setting and study period, iii) vaccine-related details, including vaccine type, source of vaccination data and the definition of ZD where the study focused on ZD, iv) types and sources of covariates, v) modelling approaches, including model type, software used, and whether modelling code was available), vi) results, such as spatial and temporal resolution of estimates and spatial aggregation units, and vii) limitations and recommendations reported by the authors. A summary of the variables we extracted from the articles is presented in Supplementary file 1 table S3. Data Synthesis Data synthesis was performed independently by two reviewers (AMN and PMM) using integrated narrative and thematic approaches to synthesize the extracted data, as these methods are well suited to integrate diverse findings from scoping reviews. We summarized the studies by their geographical distribution, year of publication, whether they provided spatial modelling methods, spatial data or both, the type of vaccines they assessed, the target age groups used to compute coverage and how ZD was defined. Thematic summarization involved inductively identifying recurring themes related to geospatial data and modelling approaches. Modelling methods were hierarchically organized into bespoke or ad-hoc themes and sub-themes based on the core concepts and relative methodological complexity. Covariates identified from the studies were thematically classified into broad categories (such as demographic, and health system) and specific sub-categories like household characteristics factors. This classification was informed primarily by the contextual usage of the covariates in the studies. Modelled estimates were summarized based on the aggregation units (for example, by administrative region or by pixel), spatial and temporal resolution, while tools, packages, reported limitations and recommendations were presented narratively. In addition, we identified limitations beyond those explicitly reported in the studies and proposed recommendations to address these gaps, with the aim of strengthening future ZD prevalence modelling. Results Article Search and Screening Of 15,587 initially retrieved articles, including 15,586 sourced from databases and one from citation searching, 7,871 duplicates were removed. The remaining 7,716 articles underwent title and abstract screening, resulting in the exclusion of 7,445 articles that did not meet the eligibility criteria (Figure 1). These 7,445 excluded articles were either literature reviews, studies focused on immunization other than routine childhood vaccination (e.g., COVID-19) or studies that did not include or describe spatial data or methods. Of the 271 articles selected for full-text review, the full text of one article could not be found. We further excluded 169 articles that did not use geospatial methods (n = 151), did not focus on childhood vaccination (n=7) or vaccination coverage (n=5), or focused on disease modelling without addressing vaccination coverage (n=5). Finally, 102 articles were included in the review (Supplementary File 2 table S1). Study Characteristics Among all 76 LLMICs (39), 68 (89.5%) countries had at least one study included in this review (Figure 2). Of the 102 studies, 18 (17.6%) were conducted across multiple countries (Supplementary file 1 table S4), while 84 (82.4%) focused on a single country (Supplementary file 1 table S5). Ethiopia (38%), Nigeria (30%), and India (19%) accounted for the largest portion of studies, together representing 87% of all included studies. Of the 84 single-country studies, 90% covered the entire country, while the remainder focused on specific subnational areas, including states, provinces, districts, counties and cities as outlined in Supplementary file 1 table S5. The included studies were published between 2010 and 2025, with the majority (70%) published between 2021 and 2024 (Figure 3). Most studies (n=86; 84.3%) provided both geospatial data and methodological approaches for modelling ZD children. The remaining 16 studies (15.7%) provided geospatial data on factors associated with immunization coverage but did not explicitly model vaccination coverage. Instead, these studies primarily examined associations or correlations between covariates and vaccination outcomes. Vaccination Vaccines assessed Unique instances of vaccines assessed across the included studies were: Bacillus Calmette–Guérin (BCG) (n=41), any dose of DTP (n=67), measles containing vaccines (MCV) (n=69), oral polio vaccine (OPV) (n=40), inactivated polio vaccine (IPV) (n=2), pneumococcal conjugate vaccine (PCV) (n=7), rotavirus vaccine (n=9), and Hepatitis B virus vaccine (HBV) (n=4). MCV1 was the most frequently studied vaccine, appearing in 66.7% (n=68) of the studies, followed by DTP1 (n=55) and DTP3 (n=55). A tabulation of vaccines and the corresponding number of studies is presented in Supplementary file 1 table S6, with some assessing more than one vaccine (n=73). Across all studies, 24 different age groups were used to report coverage of 21 vaccines (Supplementary file 1 table S20). The number of target age groups used for the same vaccine also varied, ranging from fewer than five for 13 vaccines to more than ten for four vaccines. For instance, MCV1 was reported using 21 target age groups across 67 studies, with the most common being children aged 12–23 months (n=32). Similarly, coverages for DTP1 and DTP3 were reported using ten and twelve age groups, respectively, with children aged 12–23 months being the most used age group. See Supplementary file 1 table S20 for a detailed tabulation of the age groups. ZD definition Only 19.6% of the included studies (n=20) focused on ZD prevalence (Supplementary file 1 table S7). Across these 20 studies, ZD was operationalized inconsistently, with two dominant definitions and four target age-groups, limiting comparability across settings. Specifically, a ZD child was defined as: one who had not received DTP1 in 17 studies (6,10,24,50,54–65) or one missing a single dose in the routine vaccination schedule in three studies (66–68). Four of the 20 studies only presented descriptive statistics or conducted simple analyses examining the relationship between ZD determinants (but contributed geospatial data) and ZD status (10,26,62,67) without predicting ZD. Of the 20 studies, one spanned the whole of SSA (55) and two covered all low- and middle-income countries (LMICs) (26,64). The rest were conducted mainly in Nigeria (n=8), Ethiopia (n=6), DRC (n=5), India (n=5) and Pakistan (n=4), with four studies spanning multiple countries. Over 50% of these studies (n=13) were published in 2023 or later. Five age groups were used to define the target population for evaluating ZD children, including <1 year (n=2), 0–23 months (n=1), 12–23 months (n=12), 12–35 months (n=3) and <5 years(n=1). One study did not specify the target age groups used to define ZD. The detailed summary of the target age-groups used is presented in Supplementary file 1 table S7. Data characteristics and sources Vaccination data Survey data only Of the 102 studies, 85 obtained vaccination status data from survey-based sources. These studies primarily relied on DHS data, either as a standalone source (n = 57) or in combination with other household surveys like Multiple Indicator Cluster Surveys (MICS) (n=13). Other studies (n=12) used field survey data collected after Supplementary Immunization Activities (SIA), clinical trials and community demonstration projects (Supplementary file 1 table S8). Additionally, three studies used pre-modelled vaccination estimates from the Institute for Health Metrics and Evaluation (IHME)’s local burden of disease project (26,64,69) which were derived using data from multiple household surveys. Routine data only Eleven studies conducted in Kenya, Nigeria, Ethiopia, Mozambique, Pakistan, Zambia, and Cameroon used routine data only. These studies relied on data from national health information systems such as DHIS2 and health district surveillance systems (Supplementary file 1 table S8). Studies combining survey and routine data Only five studies combined routine and survey data to analyse vaccine coverage. In Ethiopia, two studies integrated DHIS2 and DHS data. In Zambia, one study triangulated routine Health Management Information System (HMIS) data with data from 20 household surveys, including DHS and MICS. In Pakistan, one study combined health facility records with survey data collected after SIA. In Madagascar, one study used vaccination data from health centre EPI records alongside data from a longitudinal cohort study conducted in a single district. More details on the sources of vaccination data are outlined in Supplementary file 1 tables S8 & S9. Covariates Types of covariates We extracted details on covariates (factors, determinants or variables associated with ZD prevalence or low immunization coverage) from 81 of the 102 included studies. Of the remaining 21 studies, five modelled vaccination coverage without including covariates (64,65,70–72), while 15 did not predict vaccine coverage and thus they did not provide covariates (Supplementary file 1 table S10). The number and types of covariates used varied substantially across studies, with 15 studies using fewer than six covariates and 41 studies including ten or more. Since individual studies had multiple covariates, the percentages reported in this section are based on the total number (n=931) of covariates (unique instances) extracted and not the number of studies. There were six general categories of covariates used (Figure 4). Demographic factors, accounted for nearly half (49.4%) of all the covariates used, while health system factors and physical and environmental factors represented 19.2% and 15.9% of all covariates, respectively. Factors representing hard-to-reach contexts, such as exposure to conflict, and remoteness comprised only 8.1% while geographic region, infrastructural factors, such as access to electricity and distance to transport networks, accounted for just 3.9%. Other less frequently used covariates represented spatial characteristics such as the region where the child resided, the geographical size of the region or the health area. These were classified as “others” and accounted for 4% of all covariates. In terms of the number of studies using specific categories of covariates, 95% included demographic covariates, 76% included at least one health system–related factor, and only 40% included a factor representing marginalization or hard-to-reach populations (Supplementary file 1 table S11). The distribution of covariates used in ZD studies only was largely consistent with that observed across all the 102 studies, with a slightly greater representation of variables capturing hard-to-reach populations (Supplementary file 1 figure S1). A further breakdown of covariates beyond the six classes (Table 1) shows that the most frequently used variables are those linked to healthcare access and utilization, education and awareness, socio-economic status, characteristics of the child and the household and environmental factors. Health financing and covariates representing hardship areas such as presence of conflicts and slums, migration were featured the least. Table 1. Summary of covariates and data inputs across included studies Broad category Subcategory Specific aspects Source of the geospatial data Demographic (n=459, 49.3%) Population (31) Density (22), subset of the population (4) and other characteristics (5) Empirically modelled (1), Census (3), NGO, census, HMIS, UN agency (1), Global mapping projects (24), Existing survey (1) Urbanicity (50) Urban/rural dichotomy (39), urban continuum (7) and settlements (4) Global mapping projects (5), UN agency (1), Existing survey (31), European Commission (6), Satellite-based (1), University research unit (1), Primary data collection (1), Electronic Immunization Registry (1), Space agency (1) Age (44) Age of the care giver (13) and the child (31) Existing survey (37), Primary data collection (6) Socio-economic (77) Household wealth/assets (37), occupation (15), poverty (15), economic environment (7) and bank account (3) Empirically modelled (2), Existing survey (59), University research unit (1), Primary data collection (3), UN agency (2), Global mapping projects (8) Education and awareness (91) Maternal education (42), literacy (12), Media and internet (23), knowledge of disease (4) and other aspects of education (10) Existing survey (73), Global mapping projects (4), Primary data collection (8), Electronic Immunization Registry (1), HMIS (1), Empirically modelled (1) Culture and beliefs (34) Ethnicity (10), religion (19) and social group (5) Existing survey (27), University research unit (3), Primary data collection (3), HMIS (1) Child characteristics (53) Child's gender (28), birth order (17) and birth interval (2) and others (6) Existing survey (45), Primary data collection (5), Electronic Immunization Registry (1), HMIS (2) Household characteristics (54) Household head (17), size (11), marital status (13), number of children (10), and others (3) Existing survey (46), Primary data collection (7), Empirically modelled (1) Fertility and reproduction (14) Fertility 3), parity (6), pregnancy (4) and desire for last child (1) Existing survey (11), Global mapping projects (1), Primary data collection (2) Nutrition and hygiene (11) Water (3), sanitation (3), breast feeding and diet (5) Existing survey (9), Empirically modelled (2) Health system (n=179, 19.2%) Disease & morbidity (12) Childhood diarrhoea (1), respiratory infections (1), malnutrition (2), malaria (4), Polio 2 and others (2) Global mapping projects (6), Existing survey (1), Empirically modelled (2), Primary data collection (1) Health care utilization (91) Antenatal care (30), postnatal care (12), place of delivery (23), mode of delivery (3), skilled birth attendance (6), autonomy in decision making (8) and possession of a health card or document (9) Existing survey (83), Primary data collection (3), Global mapping projects (1), Electronic Immunization Registry (1) Health intervention (20) Bed nets (3), family planning (2) and other vaccinations (15) Existing survey (12), Primary data collection (1), Electronic Immunization Registry (1), Empirically modelled (1) Access to health care services (46) Proximity to healthcare (32), availability of healthcare facilities and workers (14) Global mapping projects (13), Primary data collection (9), Existing survey (2), Empirically modelled (13), HMIS (1), HMIS, UN agency (1), Crowdsourced (1), census (1), Satellite-based (1), UN agency (1), Electronic Immunization Registry (1) Health financing (10) Health insurance (6) and expenditure (4) Existing survey (9), Global mapping projects (1) Infrastructural (n=36, 3.9%) Access to infrastructure (36) Transport infrastructures ( 18), electricity (3) and night-time lights as a proxy (15) Crowdsourced (8), Global mapping projects (4), Empirically modelled (3), European Commission (2), Existing survey (2), Satellite-based (12), Global climate database (1), Other global databases (roads, boundaries, water bodies) (2) Hard to reach (n=72, 7.7%) Conflict (11) Conflict areas and proximity to such areas University research unit (1), Global mapping projects (9) Remoteness (56) Access to cities (20) and settlements (16) proximity to cultivated areas (6), protected areas (4) and water bodies (10) Crowdsourced (9), Satellite-based (1), Empirically modelled (5), Global mapping projects (24), Space agency (8), European Commission (3), Other global databases (roads, boundaries, water bodies) (4), Regional mapping agency (1) Migration (3) Migration status (3) Existing survey (1), Primary data collection (1), HMIS (1) Slums (2) Slum areas (2) Global health organization (1), Existing survey (1) Physical and environmental (n=148, 15.9%) Topographical (21) Elevation (16) and slope (5) Satellite-based (14), Regional mapping agency (1), National mapping agency (3), Global mapping projects (1) Hydrological (41) Precipitation (21), evapotranspiration (8), aridity (5) and soil/vegetation moisture (7) Global research partnership (7), University research unit (14), Global climate database (9), Satellite-based (7), Empirically modelled (1) Atmospheric conditions (45) Temperature (37), cloud cover percentage (3), frost day frequency (3) and mean vapour pressure (2) University research unit (17), Satellite-based (16), Global climate database (7) Vegetation conditions (21) Enhanced vegetation index (13), normalized difference vegetation index (5) and other indices (3) Satellite-based (20) Agricultural (20) Livestock density (13), irrigation ( 3) and the growing season and suitability (4) Global mapping projects (10), Primary data collection (1), UN agency (5), University research unit, UN agency (2), University research unit (1) Others (n=37, 4%) Geographical location (37) Representation of space through coordinates (1), subnational (23) or national regions (13). Satellite-based (1), Other global databases (roads, boundaries, water bodies) (7), Existing survey (21), HMIS (1), Crowdsourced (1), Global health organization (1), Primary data collection (1), Empirically modelled (1) Footnote: Values in parentheses represent the number of studies (n) Sources of data on covariates Covariate data were drawn from a wide range of sources. 51% out of all extracted covariates were sourced from secondary survey data. Other major sources included, university-led global mapping initiatives (15.9%), institutional datasets from mapping agencies and regional or global organizations (8.5%), satellite-derived products (7.8%), and primary surveys (5.6%). Other sources included national censuses, crowd-sourced data, model-derived outputs, and data from HMIS. There were a few instances (n=69 covariates) in which studies triangulated multiple data sources and overall, data sources were not reported for 39 covariates. Among the covariates obtained from secondary surveys, DHS was the most common source (88%) while MICS accounted for about 10%. Global mapping projects led by universities or independent research groups provided substantial sources of covariates: Worldpop, University of Southampton (33%) and Malaria Atlas Project, Kids Research Institute Australia and Curtin University (19%), and other similar projects such as IHME’s local and global burden of disease, Climate Research unit at University of East Anglia, Gridded Livestock of the World by Food and Agriculture Organization of the United Nations (FAO), Armed Conflict Location & Event Data. Global institutional sources included the European Commission, UN agencies, space agencies, and global partnerships. Additionally, the Moderate Resolution Imaging Spectroradiometer (MODIS) and the National Oceanic and Atmospheric Administration (NOAA) provided most satellite-based datasets. A summary of the sources of covariates is provided in Table 1 and Supplementary file 1 table S12. Pre-processing of covariates Some covariates were available as gridded surfaces which needed little to no pre-processing before being incorporated into geospatial models. However, some were derived from point data and required pre-processing to transform them into continuous surfaces. These included modelling of travel time/distance to service points using spatial accessibility models, estimating exposure to conflict using kernel density estimation approach, spatial interpolation, georeferencing, data quality, completeness and consistency checks, and subsequent imputations. Other pre-processing steps included harmonizing spatial resolution and coordinate reference systems, resampling or aggregating gridded data, and aligning covariates to consistent administrative boundaries. Some studies also screened covariates for multicollinearity and performed variable selection or dimensionality reduction prior to modelling. Modelling vaccination coverage Studies clustered around ten broad themes. Most focused on geospatial mapping of vaccination coverage, encompassing the modelling and visualisation of hotspots, under-immunisation, ZD children, and the determinants of spatial inequity. Vaccine timeliness, dropout, and their geographic drivers formed a related area of inquiry. Access to vaccines was examined through multiple proxies, with geographic accessibility and health system reach receiving particular attention in remote, fragile, and conflict-affected settings. Further themes included health system performance, delivery strategies, and the spatial integration of immunisation within broader health programmes. A further area of work addressed data quality and coverage estimation challenges using advanced spatial-statistical methods and machine learning (Supplementary file 2 table S2). Modelling methods Studies modelling vaccination coverage (n=86; 92.2%), including ZD prevalence, used a variety of methods ranging from simple spatial clustering approaches to complex approaches such as Bayesian geostatistical methods and ML. These approaches were either spatial (67.4%) or spatiotemporal (32.6%) and were summarized into five categories: i) spatial autocorrelation and clustering approaches, ii) small area estimation (SAE) methods, iii) spatial interpolation techniques, iv) ML methods and v) other methods such as disaggregation and multi-criteria decision analysis (Table 2). In many instances (n=49), studies applied several methods. Spatial autocorrelation & clustering Over half of the methodological studies (n=46, 53.5%) focused on assessing spatial autocorrelation and clustering of immunization status either as standalone analysis (n=15, 32.6%) or in combination with other spatial methods (n=31, 67.4%). Global spatial autocorrelation (using global Moran’s I) tests the overall patterns of dispersion and correlation, while local spatial autocorrelation (local Moran’s I) identifies local clusters of low or high immunization coverage. Further, hotspot analysis with the Getis-Ord Gi* Statistic is used to determine statistically significant hotspots and cold spots of vaccine coverage, showing clusters of high or low immunization coverage. Kulldorff’s spatial scan, on the other hand, employs a moving window across the study area to detect statistically significant clustering of areas with higher or lower vaccination coverage than expected. Table 2. Spatial and spatiotemporal methods used to model or map vaccination coverage General category Sub-category Method Reference Spatial autocorrelation, hot spot analysis and cluster detection (n=46) Global spatial autocorrelation Global Moran’s I (n=37) (59,66,68,71,73–105) Local spatial autocorrelation Local Moran’s I (n=15) (73,75,77,78,82,84–86,88,96,102,103,106–109) Subtype not specified (n=2) (71,100) Getis-Ord Gi* statistic (n=21) (59,66,73,76,80–83,88,90–94,97,101,109–111) Cluster detection Kulldorff’s spatial scan (n=21) (63,66,68,73,74,76–78,80,82,83,87,88,91,92,94,97,101,104,109,112,113) Small area estimation (n=11) Spatial model Various SAE spatial specifications (n=5) (65,98,107,114,115) Spatiotemporal model Various spatiotemporal SAE model specifications (n=6) (116–121) Spatial Interpolation (n=39) Other Inverse distance weighting (n=1) (59) Exact method not stated (105) Kriging Universal kriging (n=2) (82,91) Empirical Bayesian kriging (n=3) (78,82,88) Simple ordinary (n=1) (88) Ordinary kriging (n=10) (68,73,76,77,82,83,88,92,94,109) Type not specified (n=6) (66,87,89,95,97,101) Geostatistical models Bayesian model-based geostatistics (MBG) (n=21) (6,7,24,25,50,54,55,58,61,67,72,122–132) Reused outputs derived from a geostatistical model (n=1) (64) Artificial Intelligence (Machine learning (n=7)) Additive Generalized Additive Models (GAMs) (n=5) (55,61,122,133,134) Tree-based models Gradient Boosting Model (GBM), and Boosted regression trees (BRT), Random forests (RF), Decision Tree (DT), (n=6) (55,61,69,89,122,134) Regularization- linear models Least Absolute Shrinkage and Selection Operator (LASSO), Ridge classifier (n=3) (55,61,69) Neural Networks Multi-Layer Perceptron (MLP) (n=1) (56) Instance-based Nearest Neighbour (n=1) (56) Others (n=4) Disaggregation Spatial regression (n=1) (135) Multi-criteria decision analysis Vulnerability index (n=2) (6,60) Adjusted numerator (routine) and denominator (population) (n=2) (70,136) Small Area Estimation (SAE) SAE modelling frameworks were used in only 11 studies (12.8%) of all modelling studies. These spatial-statistical approaches were used to generate estimates of vaccination coverage for small geographic areas with limited sample sizes (unstable direct estimates) by borrowing strength from neighbouring areas or through use of auxiliary data such as malnutrition, incidence of VPDs like polio, among others. There was marked variation in model specification across SAE studies, including the use of spatial versus spatiotemporal frameworks. Studies also differed in whether they included covariates or not, how they modelled spatial random effects – both unstructured or/and structured components (e.g., BYM and Leroux models and definition of neighbourhood matrices), incorporation of temporal effects (e.g. autoregressive models or random walks of first or second order), in the spatiotemporal interaction terms they used (such as Type 1), and in the priors they used (e.g. inverse Gamma, Penalized Complexity). Spatial Interpolation Spatial interpolation techniques ranged from simple inverse distance weighting (n=1) and kriging (n=18), to more complex geostatistical modelling (n=22). Kriging – a spatial interpolation method that estimates values at unobserved locations by weighting nearby observations using distance-based autocorrelation – was implemented in four forms: simple, ordinary, universal, and empirical Bayesian. However, none of these studies incorporated covariates. Kriging studies were implemented mainly in ArcGIS (Esri, Redlands, CA, USA) software without the incorporation of covariates, and in most cases the resolution of these predictions was not reported. This analysis was also purely spatial with no temporal component, and the vast majority (83.3%) were conducted in Ethiopia only. Bayesian geostatistical studies interpolate coverage spatially or spatiotemporally, at 1 km or 5 km resolution. These studies employed covariates (n=21), and they were implemented primarily using R-INLA package (n = 20) and Markov Chain Monte Carlo approaches (n=2). These studies typically assumed an independent, identically distributed non-spatial error term (n=9), a Matern covariance function (n=18), a Euclidean distance function (n=8), and/or defined the spatial decay (distance beyond which there is negligible spatial correlation) using an exponential function (n=5). Most also reported validation metrics, including information criteria and cross-validation statistics. A key advantage of Bayesian geostatistical methods is their ability to quantify the uncertainty associated with predictions. Most of these studies assessed uncertainty using 95% credible intervals, standard deviations, or non-exceedance probabilities. Most MBG studies were conducted in Nigeria (n=14) and only three in Ethiopia. Machine learning Only seven studies (8%) used ML algorithms to estimate vaccination coverage. These were categorized as additive models (Generalized Additive Models), which combine linear and non-linear smoothing functions between predictors and the outcome, allowing flexible modelling of complex relationships; tree-based models (Random forests, Decision Tree, Gradient Boosting Model, Boosted regression trees), which recursively construct decision trees from input features, with each branch representing a decision that leads to a final prediction; regularization models (Least Absolute Shrinkage and Selection Operator -LASSO, ridge regression), which introduce constraints to prevent overfitting and improve generalizability of the model; neural networks (Multi-Layer Perceptron), which learn complex patterns through interconnected layers of nodes; and instance-based models (nearest neighbour), which rely on specific instances in the training data to make predictions (Table 2). Tools and reproducibility To model or map vaccination coverage, R (137) and STATA (138) were the two dominant statistical software used in 49 and 39 studies, respectively (Supplementary file 1 table S14). Desktop GIS software and specialized mapping tools were also employed with ArcGIS (Esri, Redlands, CA, USA) (n=38) and QGIS (n=6) ( (139) commonly used. These tools were primarily used for interpolation, data management and visualization of vaccination coverage data. Supplementary file 1 table S13 shows a summary of the tools and software used in the studies. Notably, to support reproducibility and transparency of the findings, only seven studies provided links to the code used to model vaccination coverage. Modelled outputs, limitations and recommendations Aggregation units, spatial and temporal resolution Of the 86 modelling studies, 26 conducted spatial autocorrelation and clustering analysis without estimating coverage, hence they did not provide any spatial resolution of outputs. Across the remaining 60 studies that modelled vaccine coverage, spatial resolution of the outputs was presented as gridded surfaces only (high resolution) (n=7), at administrative units only (n=31) or as both gridded surfaces and at administrative units (n=22). Studies providing estimates as gridded surfaces used spatial resolutions of 1km (n=16), 5km (n=9), and 10km (one multi-country study), while two studies produced estimates at both 1km and 5km. One study did not report the spatial resolution. Overall, these studies predominantly employed geostatistical methods and covered 11 countries. The 22 studies that aggregated estimates to administrative units did so at varying levels of granularity: the first administrative level (region, provinces or states; n=4), at the second administrative unit (districts, counties or local government areas (LGAs) n=16) or at finer units such as cities in Niger and wards in Nigeria (Supplementary file 1 tables S15 & S16). Regarding temporal resolution, of the 86 modelling studies, 27 produced annual estimates comprising of 18 country-level, 8 multi-country and 1 subnational level study. The remaining 59 studies reported estimates for a single point only (year). Reported limitations Data limitations Issues related to vaccination data quality dominated the reported limitations. Since most studies relied on survey-based vaccination data, data quality limitations were mainly due to recall bias, non-response bias, and displacement of GPS coordinates. Studies using routine administrative data instead highlighted inaccuracies in the population denominators (the target population of children aged under five years who are due or eligible for a specific vaccine according to the national immunization schedule) used to estimate vaccination coverage. Data scarcity, incompleteness, or unavailability, particularly in hard-to-reach or conflict-affected areas, was another major challenge, directly affecting the representativeness of vaccination data. Additional constraints included the low spatial resolution of survey data due to aggregation at administrative levels (e.g., districts), the use of outdated vaccination data, cross-section nature of survey data and the inability to account for immunization delivered through SIAs. Challenges with covariate data were mainly exclusion of key contextual covariates, such as supply-side factors like vaccine stocks, conflict exposure and migration in modelling, due to their unavailability or outdatedness. Additionally, use of covariates from multiple sources caused difficulties harmonizing multiple data sources. Methodological limitations Uncertainty in coverage estimates was not consistently reported, as some studies reported point estimates without associated confidence or credible intervals, limiting interpretation of the estimates. The ability to assess temporal patterns in coverage was also restricted because most analyses relied on single survey rounds or lacked longitudinal vaccination data. Studies that aggregated results to higher administrative units (e.g., provinces, districts) noted that this may have masked subnational heterogeneities. Some limitations were specific to the methods used, for instance, kriging performed poorly in unsampled areas, while SaTScan – which relies on circular scanning windows – may have missed irregularly shaped clusters. Studies using travel-time-based accessibility models cited oversimplification of real-world conditions by not accounting for weather conditions, transport availability, or traffic congestion. More details of data and methodological limitations are provided in Supplementary file 1 table S17. Reported recommendations Studies emphasized improving vaccination data by using up-to-date datasets (including administrative boundaries, health facilities, vaccination status, and population distribution) and incorporating additional covariates such as vaccine hesitancy, facility readiness, and care-seeking behaviour. Standardized data collection guidelines, rigorous data quality checks, and triangulation of multiple sources – including census data, satellite-derived population estimates, and age-stratified population data – were recommended to enhance accuracy, account for seasonal migration, and reduce uncertainties in coverage estimates. Integrating data from multiple cross-sectional surveys and including qualitative information was also suggested to better help interpret coverage patterns, while strengthening the completeness and quality of routine immunization data was highlighted as critical for continuous monitoring, particularly at the subnational level. Methodological considerations to improve vaccination coverage modelling included subnational analyses to capture local heterogeneities, provision of model uncertainty to enhance interpretability and reliability of estimates, and continuous model refinement to include missing covariates and address geographic or urban-rural variability. Use of joint modelling frameworks to triangulate routine and survey vaccination data, as well as ML methods for automated feature extraction were also recommended. In addition, some studies proposed using online dashboards to facilitate interpretation and dissemination of results. Finally, continuous monitoring and evaluation of vaccination coverage, particularly in low-coverage areas, was recommended to track trends, inform interventions, and support adaptive modelling strategies over time. See Supplementary file 1 table S18 & S19 for more details. Discussion Despite concerted efforts aimed at achieving full childhood immunization by key actors like GAVI and UNICEF (20,21,140), pockets of ZD children persist in LLMICs (141). Effectively reaching and vaccinating these children requires their accurate identification using spatially resolved data and robust geospatial methods. Our scoping review identified 102 studies conducted across 68 LLMICs that outlined relevant spatial data, spatial or spatial-statistical methods used to estimate childhood vaccine coverage, and the associated gaps. Nine in ten studies were conducted in just three countries (Ethiopia, Nigeria, or India), likely due to the high ZD burden in these countries (2). Additionally, 70% were published between 2021 and 2025, indicating that much of the evidence has emerged only recently. Regardless of the growing recognition of the value of geospatial approaches for immunisation programming, our review shows a complex and uneven landscape. The mapping of ZD children, while gaining traction, remains limited in scope and uptake relative to its demonstrated utility for identifying pockets of under-immunisation. Definitions of ZD vary considerably across studies, undermining comparability and policy utility of findings. Data sources are heavily skewed toward household surveys, with routine administrative data remaining largely underused despite its potential for continuous, granular monitoring. Most studies rely on simpler exploratory methods without uncertainty-quantified estimates that operational microplanning demands, and spatiotemporal frameworks remain underused, limiting the capacity to track vaccination trends and link fluctuations to programmatic drivers. Below, we discuss these each point and its implications for the identification of ZD children in LLMICs. Only a fifth of the studies focused on mapping ZD prevalence indicating limited uptake of geospatial approaches for targeting ZD children, despite growing evidence of their ability to detect small pockets of under-immunization (50,55). This could be due to technical and structural barriers; ZD clusters tend to be rare, spatially heterogeneous, and often located in marginalized settings (26,50). Identifying them requires fine-resolution spatial data, advanced analyses, and cross-sector collaboration, which remain unevenly available in many LLMICs. Due to this complexity, such modelling studies are led by a handful of well-resourced academic groups, e.g., IHME local burden of disease and Worldpop’s VaxPop project (137,138) . Limited translation of ZD estimates in routine country planning, where administrative unit averages still dominate may further discourage adoption of geospatial approaches (49,144) Across the studies, ZD was conceptualised as: i) not receiving any vaccine (66–68), or ii) missing DTP1 (6,10,24,50,54–65). Inconsistent definitions of ZD capture different populations which may lead to over or under-estimation of ZD prevalence and subsequently, important consequences on policy and programmatic actions such as inefficient strategies for reaching the most disadvantaged children (130). For example, globally, 14.2% of children are DTP1-ZD and 7.5% are completely ZD (no single vaccine) (145). Despite recommended use of DTP1 for operational purposes (4), our review emphasizes the need to further standardize and harmonise definitions of ZD, and appropriate target age-groups. Inconsistent definitions and target age-groups lead to incomparability of evidence limiting its usefulness for global policy formulation, prioritization and benchmarking progress. In addition, inconsistent definitions can yield divergent interpretations of drivers of ZD and weaken assessments of equity gradients across settings. On the other hand, definitional variation is not without rationale. DTP1 definition carries clear operational justification: administrative data capture DTP1 non-receipt continuously and at facility level, enabling real-time monitoring at a granularity that survey-based definitions cannot match, and at population level DTP1 coverage tracks closely with rates of complete non-vaccination (146,147). Conversely, the broader definition of receiving no vaccine may better capture the most marginalised children in settings where even partial immunisation is rare (146,147). The main source of both vaccination and covariate data was household survey data, particularly DHS ( Supplementary file 1 table S8 ). These surveys are nationally representative, standardized, less vulnerable to reporting incompleteness or treatment-seeking biases with well-defined population denominators and detailed socio-demographic indicators. However, they do not capture subnational heterogeneities due to low spatial resolution (missing most vulnerable populations), are cross-sectional, and are conducted less frequently – every three to five years (148). The recent defunding of the DHS Program further threatens the tracking of health indicators in LLMICs including ZD (35). In contrast, there was very low use of routine administrative vaccination data despite its increased availability in the last decade through DHIS2 (149). Routine data is continuous, timely and geographically disaggregated and can enable continuous ZD monitoring and localized decision-making. However, it is prone to incompleteness, outliers, and inaccurate population denominators for coverage computation (62,112,150). Further, as routine data are collected at the health facility level, estimating denominators and health facility catchment areas is challenging (151), particularly in settings with high population mobility, inaccurate population estimates and low healthcare utilization rates. Therefore, making use of routine data requires advanced methods, which may explain why studies preferred survey-based data (114). This highlights a need to address systemic and structural gaps to improve collection and reporting of HMIS data (152), use of rapid and targeted assessment tools such as lot quality assurance sampling and micro-planning to pinpoint low coverage areas (153), and triangulation of survey and routine data to leverage their differing strengths (151). Only five studies used both survey and routine data, treating them separately (e.g., using DHS coverage to adjust DHIS2 denominators) rather than jointly modelling them. There is need for joint integration of routine and survey data in a modelling framework to maximise the complementary strengths of both data types. Such integrated frameworks have proven superior in other applications, for instance in mapping malaria incidence (151), by improving estimation accuracy and revealing hidden inequities that single-source models miss. Establishing integrated routine–survey modelling pipelines could represent a high-impact priority for the next generation of ZD mapping, enabling both frequent monitoring and robust estimation in data-sparse areas. We identified six categories of the covariates used, with demographic, health-system, and physical-environment factors being the most applied, consistent with previous findings that highlight demographic and healthcare access factors as key barriers to vaccination among ZD communities (154). Still, there is a significant gap in the covariates used to capture fragile, marginalized, or hard-to-reach populations – such as conflict-affected, remote, or highly mobile populations (e.g., nomadic pastoralists, refugees, internally displaced persons), and urban slum dwellers. This was mainly due to data unavailability or scarcity in these settings (26,54,55). Given that ZD children are disproportionately concentrated in marginalized and fragile settings, the lack of covariates capturing these contexts limits the ability to identify populations at greatest risk of ZD. Studies modelling ZD prevalence should therefore incorporate at least one covariate reflecting marginalization, depending on the study context and scope. Collaborative efforts by data agencies, local health authorities and international partners are needed to expand covariate data availability to avoid systematically missing the most disadvantaged populations. While there were diverse modelling methods, most studies relied on simpler, exploratory approaches to examine autocorrelation and clustering of vaccination data. These methods do not interpolate estimates for unsampled locations and thus cannot generate gridded estimates. Gridded or continuous estimates are preferable, as they can be directly overlaid with high resolution population density maps for absolute estimates of ZD or other gridded estimate to identify overlapping dual or triple burdens (6,26). These exploratory approaches were implemented in mapping software (especially ArcGIS (Esri, Redlands, CA, USA)) which have a graphical user interface and require less technical expertise and computational resources than advanced modelling approaches. However, these descriptive clustering approaches alone are insufficient for operational microplanning, which increasingly requires small area estimates with quantified uncertainty for prioritization and monitoring. More complex modelling approaches including SAE, MBG and ML are better suited to ZD estimation as they incorporate covariates, generate uncertainty and produce outputs at operationally meaningful scales (155,156). The resolution at which the modelled outputs are produced matters. Gridded estimates are particularly valuable for hotspot identification and the targeting of interventions, as they can be directly overlaid with high-resolution population data to identify absolute numbers of ZD children and areas of overlapping deprivation (67,125). These gridded outputs can then be aggregated to subnational administrative units that align with the scales at which immunisation programmes are planned and evaluated (114,125), support decentralised resource targeting, and reveal subnational inequities masked by national averages (49). Aggregation and produces outputs more interpretable for health managers, while enabling alignment with global policy benchmarks such as GVAP, IA2030) (122). Yet, uptake of these methods remains low in low-resource settings due to steep requirements for expertise and computing resources. National programs in LLMICs often lack sufficient local capacity and often outsource complex analyses, reducing ownership and timeliness evidence for decision-making (157). Compounding this, very few studies provided analytical code, limiting reproducibility and the ability of country teams to adapt and reuse methods. Alternative user-friendly web tools such as Maplaria (158), MBGapp (159), and sae4health (160) address these barriers by automating pipelines, hosting computations remotely, and requiring no coding or advanced statistical knowledge. Bridging this gap through investment in local capacity, user-friendly analytical tools, and stronger partnerships between academic groups and national programmes is essential to strengthen country ownership of geospatial evidence and improve the timeliness of targeting decisions for zero-dose programming. Few studies used AI approaches, indicative of a slow uptake of these methods in this domain. While a comparative study showed that geostatistical models slightly outperform ML models in estimating vaccine coverage (29), it is still necessary to explore the full potential of AI approaches, especially as the availability of geospatial big data and satellite-derived covariates increases. These models have the capacity to handle large, multidimensional datasets, such as DHS and routine HMIS data, enabling integration of variables (e.g., demographic, health system, environmental) and recognition of non-linear patterns in the variables (47). They can also improve the efficiency of analytical tasks, for instance, using Large Language Models (LLMs) to automate feature extraction (161,162). Only 31% of the studies incorporated a spatiotemporal framework, reflecting the complexity of spatiotemporal modelling and the scarcity of longitudinal data, as most analyses relied on cross-sectional survey data. Temporal estimates are critical for understanding vaccination trends over time and linking fluctuations to events or seasonal factors. Where longitudinal routine data are available, spatiotemporal approaches bridge microplanning needs and dynamic monitoring of immunization performance. Promoting the availability and use of longitudinal routine data, alongside advancing spatiotemporal modelling approaches is essential to enhance temporal analysis and prediction of ZD prevalence. The role of infection prevalence, transmission dynamics, and disease burden remains insufficiently integrated into studies assessing correlations with ZD children, potentially biasing interpretation of observed associations. To advance this field, our review characterises the landscape of spatial vaccine coverage modelling in LLMICs, highlighting persistent data and methodological limitations that constrain the accuracy, equity, and comparability of estimates. Hard-to-reach populations are systematically excluded from analysis due to the lack of both reliable vaccination data and covariates capturing key contextual factors such as conflict exposure, poverty, remoteness and migration. This leads to model misspecification in the settings where ZD burden is highest. Heavy reliance on household surveys, with limited triangulation with DHIS2 data, further restricts the effective use of routine data, which are themselves undermined by incomplete reporting and inaccurate population denominators. Addressing these challenges will require harmonised definitions of ZD children, improved integration of survey and routine data through joint modelling frameworks, and the incorporation of non-traditional data sources such as satellite imagery, mobility data, and conflict databases. It will also require strengthening of data collection in underserved settings through approaches such as community health worker enumeration and rapid assessments ( Supplementary File 3) . Strengths and Limitations This review synthesizes evidence from a broad range of geospatial studies focused on assessing childhood vaccination coverage, providing a consolidated understanding the gaps in data and sources, methods, and structural limitations of studies already available in the literature. To our knowledge, it is the first review to examine ZD prevalence through a geospatial lens, offering a unique contribution to the field and helping clarify a potential roadmap on how spatial data and methods can be applied to identify ZD children. However, the review only included articles published in English, which may have resulted in the exclusion of relevant studies published in other languages. In addition, the included studies varied in ZD definitions, denominators, spatial scales, and modelling approaches, which limited direct comparability of estimates across settings. Conclusion Our review emphasizes the critical role that geospatial data and methods play in modelling and mapping ZD prevalence across diverse settings. It also highlights substantial gaps and limitations in the data and methods used, offering insights into where improvements are needed to better identify ZD children. Future research should prioritize triangulating routine immunization and household survey data while placing greater emphasis on understudied populations such as nomadic groups, refugees, urban slum residents, and conflict-prone areas. In addition, studies should move beyond describing spatial patterns through clustering and similar methods and progress toward generating monitoring tools and predictive estimates of ZD prevalence at detailed spatiotemporal scales, which are more actionable for targeted interventions and policymaking. Overall, this review provides a foundational reference for future ZD surveillance and modelling work as well as for policymakers seeking to integrate geospatial evidence into programming and monitoring. The covariates and methodological approaches identified across studies offer a practical starting point for researchers, while the documented gaps and recommendations can help guide efforts to strengthen, harmonize, and streamline geospatial modelling of ZD populations. Ultimately, improving standardization, strengthening routine data and analytic capacity, and expanding research into fragile and underserved contexts will be essential to ensure that geospatial evidence translates into equitable immunization gains. Declarations The REACHOUT Consortium Carlo Federici, Samuel Muhula, Jeanine Condo, Fabrizio Tediosi, Bolanle Oyeledun, Piero Poletti, Francesco Menegale, Manuela De Allegri, John Kutna, Yvonne Opanga, Herbert Barasa, Joan Mboga, Anne Gitimu, Caroline Mudereri, Gashaija Absolomon, Felix Rubuga Kitema, Hassan Sibomana, Olivier Wane, Jean Damascene Hagenimana, Grace Kabanyana, Jean de Dieu Hakizimana, Emma Clarke-Deelder, Amit Aryal, Obioma Ezebuka, Francis Ogirima, Oluwatosin Oladokun, Abimbola Phillips, David Udanwojo, Michael Omobhude, Carlos Felipe Balmaceda, Alessia Melegaro, Maria Cucinello, Aleksandra Torbica, Vittoria Offeddu, Ankit Shanker, Phidelis Wamalwa, Kavita Singh, Swati Srivastava CRediT authorship contribution statement Conceptualization: PMM, AMN, CF Methodology: PMM, AMN Software: PMM, AMN, MMM Validation: PMM, AMN Formal analysis: PMM, AMN, MMM Investigation: PMM, AMN Resources: PMM Data Curation: PMM, AMN, MMM Writing - Original Draft: PMM, AMN Writing - Review & Editing, PMM, AMN, MMM, SS, CM, FRK, YO, EC, FM, AM, AG, JC, FK, LB, JIB, PP, AT, CF, AM Visualization: PMM, AMN Supervision: PMM Project administration: PMM, CF Funding acquisition: PMM, SS, JC, PP, AT, AM, CF Ethical approval Not applicable. Declaration of Generative AI and AI-assisted technologies in the writing process ChatGPT- text editing Funding: The work was funded by the European Union under The Global Health EDCTP3 Joint Undertaking (GH EDCTP3 JU) – Grant Agreement number 101159477. Views and opinions expressed are however, those of the authors only and do not necessarily reflect those of the European Union or GH EDCTP3. Neither the European Union nor the granting authority can be held responsible for them . PMM is supported by Fonds voor Wetenschappelijk Onderzoek (FWO) for his Senior Postdoctoral Fellowship (Grant number: 1201925N) Declaration of competing interests The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements We thank Rehema Ouko (Department of Public Health, Institute of Tropical Medicine (ITM), Antwerp, Belgium) and Dr. Aliki Christou (Department of Public Health, Institute of Tropical Medicine (ITM), Antwerp, Belgium) for their assistance in obtaining access to relevant articles used in this review. Data Availability This study is based on secondary analysis of previously published studies. References WHO. Immunization [Internet]. [cited 2025 Nov 25]. Available from: https://www.afro.who.int/health-topics/immunization WUENIC Trends [Internet]. [cited 2026 Jan 27]. Available from: https://worldhealthorg.shinyapps.io/wuenic-trends/ GAVI. The Zero-Dose Child: Explained [Internet]. [cited 2025 Nov 25]. Available from: https://www.gavi.org/vaccineswork/zero-dose-child-explained GAVI. Zero-dose children and missed communities [Internet]. [cited 2025 Nov 25]. Available from: https://www.gavi.org/our-alliance/strategy/phase-5-2021-2025/equity-goal/zero-dose-children-missed-communities Corrêa GC, Uddin MJ, Wahed T, Oliveras E, Morgan C, Kamya MR, et al. Measuring Zero-Dose Children: Reflections on Age Cohort Flexibilities for Targeted Immunization Surveys at the Local Level. Vaccines (Basel). 2024 Feb 1;12(2). doi:10.3390/vaccines12020195 Utazi CE, Chan HMT, Olowe I, Wigley A, Tejedor-Garavito N, Cunningham A, et al. A zero-dose vulnerability index for equity assessment and spatial prioritization in low- and middle-income countries. Spatial Statistics. Elsevier B.V.; 2023. doi:10.1016/j.spasta.2023.100772 Arambepola R, Yang Y, Hutchinson K, Mwansa FD, Doherty JA, Bwalya F, et al. Using geospatial models to map zero-dose children: Factors associated with zero-dose vaccination status before and after a mass measles and rubella vaccination campaign in Southern province, Zambia. BMJ Glob Health. 2021 Dec 30;6(12). doi:10.1136/bmjgh-2021-007479 Rachlin A, Danovaro-Holliday MC, Murphy P, Sodha S V., Wallace AS. Routine Vaccination Coverage — Worldwide, 2021. MMWR Morb Mortal Wkly Rep. 2022;71(44). doi:10.15585/mmwr.mm7144a2 Utazi CE, Thorley J, Alegana VA, Ferrari MJ, Takahashi S, Metcalf CJE, et al. Mapping vaccination coverage to explore the effects of delivery mechanisms and inform vaccination strategies. Nat Commun. 2019 Dec 1;10(1). doi:10.1038/s41467-019-09611-1 PubMed PMID: 30967543. Mbunga BK, Liu PY, Bangelesa F, Mafuta E, Dalau NM, Egbende L, et al. Zero-Dose Childhood Vaccination Status in Rural Democratic Republic of Congo: Quantifying the Relative Impact of Geographic Accessibility and Attitudes toward Vaccination. Vaccines (Basel). 2024 Jun 1;12(6). doi:10.3390/vaccines12060617 Arambepola R, Yang Y, Hutchinson K, Mwansa FD, Doherty JA, Bwalya F, et al. Using geospatial models to map zero-dose children: Factors associated with zero-dose vaccination status before and after a mass measles and rubella vaccination campaign in Southern province, Zambia. BMJ Glob Health. 2021 Dec 30;6(12). doi:10.1136/bmjgh-2021-007479 FMoHSW FM of H and SW of N, NPC/Nigeria NPC, ICF. Nigeria Demographic and Health Survey 2024 [Internet]. 2025 Oct [cited 2025 Nov 25]. Available from: https://dhsprogram.com/publications/publication-FR395-DHS-Final-Reports.cfm GAVI. Turning the corner: IA2030 Global Report 2023. 2023. Murhekar M V., Kumar MS. Reaching zero-dose children in India: progress and challenges ahead. The Lancet Global Health. Elsevier Ltd; 2021. p. e1630–1. doi:10.1016/S2214-109X(21)00406-X PubMed PMID: 34798011. Malande OO, Meyer JC, Adamu VE, Munube D, Muhindo R, Kigen B, et al. Zero-dose or under immunized: Where should the emphasis lie? Journal of the African Society for Paediatric Infectious Diseases. 2024 Dec 30;3:1–9. doi:10.15641/jafspidvol3pp1-9/1732 Progress and challenges with Achieving Universal Immunization Coverage [Internet]. [cited 2025 Nov 21]. Available from: https://www.who.int/publications/m/item/progress-and-challenges Rainey JJ, Watkins M, Ryman TK, Sandhu P, Bo A, Banerjee K. Reasons related to non-vaccination and under-vaccination of children in low and middle income countries: Findings from a systematic review of the published literature, 1999-2009. Vaccine. 2011. p. 8215–21. doi:10.1016/j.vaccine.2011.08.096 PubMed PMID: 21893149. Jani J V., De Schacht C, Jani I V., Bjune G. Risk factors for incomplete vaccination and missed opportunity for immunization in rural Mozambique. BMC Public Health. 2008;8. doi:10.1186/1471-2458-8-161 PubMed PMID: 18485194. GAVI. IMMUNIZATION AGENDA 2030: A Global Strategy to Leave No One Behind [Internet]. 2020 [cited 2025 Aug 22]. Available from: https://www.who.int/teams/immunization-vaccines-and-biologicals/strategies/ia2030 GAVI. Phase 5 (2021–2025) [Internet]. [cited 2025 Nov 20]. Available from: https://www.gavi.org/our-alliance/strategy/phase-5-2021-2025 Phase 6 (2026–2030) [Internet]. [cited 2025 Nov 24]. Available from: https://www.gavi.org/our-alliance/strategy/phase-6-2026-2030#vision WHO. IA2030 IG 2.1: Number of zero dose children [Internet]. [cited 2025 Nov 25]. Available from: https://www.who.int/data/gho/indicator-metadata-registry/imr-details/7792 United Nations. The Sustainable Development Goals Report [Internet]. 2025 [cited 2025 Aug 12]. Available from: https://unstats.un.org/sdgs/report/2025/The-Sustainable-Development-Goals-Report-2025.pdf Utazi CE, Aheto JMK, Wigley A, Tejedor-Garavito N, Bonnie A, Nnanatu CC, et al. Mapping the distribution of zero-dose children to assess the performance of vaccine delivery strategies and their relationships with measles incidence in Nigeria. Vaccine. 2023 Jan 4;41(1):170–81. doi:10.1016/j.vaccine.2022.11.026 PubMed PMID: 36414476. Utazi CE, Thorley J, Alegana VA, Ferrari MJ, Takahashi S, Metcalf CJE, et al. High resolution age-structured mapping of childhood vaccination coverage in low and middle income countries. Vaccine. 2018 Mar 14;36(12):1583–91. doi:10.1016/j.vaccine.2018.02.020 PubMed PMID: 29454519. Wigley A, Lorin J, Hogan D, Utazi CE, Hagedorn B, Dansereau E, et al. Estimates of the number and distribution of zero-dose and under-immunised children across remote-rural, urban, and conflict-affected settings in low and middle-income countries. PLOS Global Public Health. 2022 Oct 1;2(10 October). doi:10.1371/journal.pgph.0001126 GAVI. IRMMA framework – guidance on use of Gavi support to reach zero-dose children and missed communities [Internet]. 2021 [cited 2025 Aug 26]. Available from: https://immunizationeconomics.org/recent-activity/2021/8/1/irmma-framework-guidance-on-use-of-gavi-support-to-reach-zero-dose-children-and-missed-communities/ Utazi CE, Thorley J, Alegana VA, Ferrari MJ, Takahashi S, Metcalf CJE, et al. High resolution age-structured mapping of childhood vaccination coverage in low and middle income countries. Vaccine. 2018 Mar 14;36(12):1583–91. doi:10.1016/j.vaccine.2018.02.020 PubMed PMID: 29454519. Utazi CE, Yankey O, Chaudhuri S, Olowe ID, Danovaro-Holliday MC, Lazar AN, et al. Geostatistical and machine learning approaches for high-resolution mapping of vaccination coverage. Spat Spatiotemporal Epidemiol. 2025 Aug 1;54:100744. doi:10.1016/J.SSTE.2025.100744 PubMed PMID: 40935503. Tatem AJ. Innovation to impact in spatial epidemiology Sally Blower; Gerardo Chowell. BMC Med. 2018 Nov 14;16(1). doi:10.1186/s12916-018-1205-5 PubMed PMID: 30424754. Kirby RS, Delmelle E, Eberth JM. Advances in spatial epidemiology and geographic information systems. Ann Epidemiol. 2017;27(1). doi:10.1016/j.annepidem.2016.12.001 Delmelle EM, Desjardins MR, Jung P, Owusu C, Lan Y, Hohl A, et al. Uncertainty in geospatial health: challenges and opportunities ahead. Annals of Epidemiology. 2022. doi:10.1016/j.annepidem.2021.10.002 Morrison CN, Mair CF, Bates L, Duncan DT, Branas CC, Bushover BR, et al. Defining Spatial Epidemiology: A Systematic Review and Re-orientation. Epidemiology. 2024. doi:10.1097/EDE.0000000000001738 Corsi DJ, Neuman M, Finlay JE, Subramanian S V. Demographic and health surveys: A profile. Int J Epidemiol. 2012 Dec;41(6):1602–13. doi:10.1093/ije/dys184 PubMed PMID: 23148108. Khaki JJ, Molenaar J, Karki S, Olal E, Straneo M, Mosuse MA, et al. When health data go dark: the importance of the DHS Program and imagining its future. BMC Med. 2025 Dec 1;23(1). doi:10.1186/s12916-025-04062-6 PubMed PMID: 40275318. Macharia PM, Beňová L, Ray N, Semaan A, Musau MM, Kipterer JK, et al. Putting health facilities on the map: a renewed call to create geolocated, comprehensive, updated, openly licensed dataset of health facilities in sub-Saharan African countries. BMC Medicine 2025 23:1. 2025 Apr 7;23(1):211-. doi:10.1186/S12916-025-04023-Z PubMed PMID: 40197252. Population Reference Bureau. The DHS Program - Evolution of The DHS Program [Internet]. [cited 2025 Nov 25]. Available from: https://dhsprogram.com/data/visualizations/Evolution-of-The-DHS-Program.cfm DHIS2. DHIS2 In Action [Internet]. [cited 2025 Nov 25]. Available from: https://dhis2.org/in-action/#map World Bank Country and Lending Groups – World Bank Data Help Desk [Internet]. [cited 2026 Feb 3]. Available from: https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups Favin M, Steinglass R, Fields R, Banerjee K, Sawhney M. Why children are not vaccinated: A review of the grey literature. International Health. 2012. p. 229–38. doi:10.1016/j.inhe.2012.07.004 PubMed PMID: 24029668. Tauil M de C, Sato APS, Waldman EA. Factors associated with incomplete or delayed vaccination across countries: A systematic review. Vaccine. Elsevier Ltd; 2016. p. 2635–43. doi:10.1016/j.vaccine.2016.04.016 PubMed PMID: 27109562. Galadima AN, Zulkefli NAM, Said SM, Ahmad N. Factors influencing childhood immunisation uptake in Africa: a systematic review. BMC Public Health. 2021 Dec 1;21(1). doi:10.1186/s12889-021-11466-5 PubMed PMID: 34320942. Bangura JB, Xiao S, Qiu D, Ouyang F, Chen L. Barriers to childhood immunization in sub-Saharan Africa: A systematic review. BMC Public Health. BioMed Central; 2020. doi:10.1186/s12889-020-09169-4 PubMed PMID: 32664849. Guzman-Holst A, DeAntonio R, Prado-Cohrs D, Juliao P. Barriers to vaccination in Latin America: A systematic literature review. Vaccine. Elsevier Ltd; 2020. p. 470–81. doi:10.1016/j.vaccine.2019.10.088 PubMed PMID: 31767469. Belt RV, Abdullah S, Mounier-Jack S, Sodha S V., Danielson N, Dadari I, et al. Improving Equity in Urban Immunization in Low- and Middle-Income Countries: A Qualitative Document Review. Vaccines. Multidisciplinary Digital Publishing Institute (MDPI); 2023. doi:10.3390/vaccines11071200 Lyons C, Nambiar D, Johns NE, Allorant A, Bergen N, Hosseinpoor AR. Inequality in Childhood Immunization Coverage: A Scoping Review of Data Sources, Analyses, and Reporting Methods. Vaccines. Multidisciplinary Digital Publishing Institute (MDPI); 2024. doi:10.3390/vaccines12080850 Crocker-Buque T, Mindra G, Duncan R, Mounier-Jack S. Immunization, urbanization and slums - A systematic review of factors and interventions. BMC Public Health. 2017 Jun 8;17(1). doi:10.1186/s12889-017-4473-7 PubMed PMID: 28595624. Musuka G, Umar A umra, Dadari I, Moyo E, Mano O, Iradukunda PG, et al. Bridging immunization gaps in Sub-Saharan Africa: a narrative review of microplanning, geospatial, and machine learning approaches to reach zero-dose children and under-immunised children. Vaccine. 2026 Apr;79:128413. doi:10.1016/j.vaccine.2026.128413 PubMed PMID: 41806761. Reaching the unreached with life-saving vaccines through data science and geospatial technologies - World | ReliefWeb [Internet]. [cited 2026 Mar 18]. Available from: https://reliefweb.int/report/world/reaching-unreached-life-saving-vaccines-through-data-science-and-geospatial-technologies Baptiste AEJ, Wagai J, Hahné S, Adeniran A, Koko RI, de Vos S, et al. High-Resolution Geospatial Mapping of Zero-Dose and Underimmunized Children Following Nigeria’s 2021 Multiple Indicator Cluster Survey/National Immunization Coverage Survey. Journal of Infectious Diseases. 2024 Jul 15;230(1):e131–8. doi:10.1093/infdis/jiad476 PubMed PMID: 39052714. Tricco AC, Lillie E, Zarin W, O’Brien KK, Colquhoun H, Levac D, et al. PRISMA extension for scoping reviews (PRISMA-ScR): Checklist and explanation. Ann Intern Med. 2018 Oct 2;169(7):467–73. doi:10.7326/M18-0850 PubMed PMID: 30178033. Njogu A, Mutinda M, Macharia P. Geospatial data and methods for mapping zero-dose children in low- and middle-income countries: A scoping literature review protocol [Internet]. 2025 May 28. doi:10.17605/OSF.IO/2Z3XV Peters MD, Godfrey CM, Mclnerney P, Soares CB, Khalil H, Parker D. The Joanna Briggs Institute Reviewers’ Manual 2015. Methodology for JBI Scoping Reviews. 2015. Alegana VA, Ticha JM, Mwenda JM, Katsande R, Gacic-Dobo M, Danovaro-Holliday MC, et al. Modelling the spatial variability and uncertainty for under-vaccination and zero-dose children in fragile settings. Sci Rep. 2024 Dec 1;14(1). doi:10.1038/s41598-024-74982-5 PubMed PMID: 39420047. Mosser JF, Gagne-Maynard W, Rao PC, Osgood-Zimmerman A, Fullman N, Graetz N, et al. Mapping diphtheria-pertussis-tetanus vaccine coverage in Africa, 2000–2016: a spatial and temporal modelling study. The Lancet. 2019 May 4;393(10183):1843–55. doi:10.1016/S0140-6736(19)30226-0 PubMed PMID: 30961907. Biswas A, Tucker J, Bauhoff S. Performance of predictive algorithms in estimating the risk of being a zero-dose child in India, Mali and Nigeria. BMJ Glob Health. 2023 Oct 11;8(10). doi:10.1136/bmjgh-2023-012836 Edson Utazi C, Pannell O, Aheto JMK, Wigley A, Tejedor-Garavito N, Wunderlich J, et al. Assessing the characteristics of un- and under-vaccinated children in low- and middle-income countries: A multi-level cross-sectional study. PLOS Global Public Health. 2022 Apr 1;2(4 April). doi:10.1371/journal.pgph.0000244 Aheto JMK, Olowe ID, Chan HMT, Ekeh A, Dieng B, Fafunmi B, et al. Geospatial Analyses of Recent Household Surveys to Assess Changes in the Distribution of Zero-Dose Children and Their Associated Factors before and during the COVID-19 Pandemic in Nigeria. Vaccines (Basel). 2023 Dec 1;11(12). doi:10.3390/vaccines11121830 Shiferie F, Gebremedhin S, Andargie G, DelPizzo F, Belete K, Fenta TG. Spatial distribution of zero-dose children in Ethiopia: evidence for a targeted intervention from a large-scale cross-sectional evaluation survey. Front Pediatr. 2024;12. doi:10.3389/fped.2024.1337922 Saidu Y, Agbor VN, Di Mattei P, Nchinjoh SC, Edwidge NN, Njoh AA, et al. Development of a composite scoring system to rank communities at high risk of zero-dose children in Cameroon: A geospatial analysis. J Glob Health. 2023;13. doi:10.7189/jogh.13.04136 PubMed PMID: 37971948. Haeuser E, Nguyen JQ, Rolfe S, Nesbit O, Fullman N, Mosser JF. Assessing Geographic Overlap between Zero-Dose Diphtheria–Tetanus–Pertussis Vaccination Prevalence and Other Health Indicators. Vaccines (Basel). 2023 Apr 1;11(4). doi:10.3390/vaccines11040802 Mehmood M, Setayesh H, Siddiqi DA, Siddique M, Iftikhar S, Soundardjee R, et al. Prevalence, geographical distribution and factors associated with pentavalent vaccine zero dose status among children in Sindh, Pakistan: analysis of data from the 2017 and 2018 birth cohorts enrolled in the provincial electronic immunisation registry. BMJ Open. 2022 May 18;12(5). doi:10.1136/bmjopen-2021-058985 PubMed PMID: 35584879. Mutwadi AM, Madinga JN, Vanlerberghe V, Mbala PK, van der Sande MAB. Geospatial and multilevel clustering of zerodose children in Kikwit, Democratic Republic of the Congo in 2022. PLOS Global Public Health. 2024 Feb 29;4(2). doi:10.1371/journal.pgph.0002617 Fullman N, Correa GC, Ikilezi G, Phillips DE, Reynolds HW. Assessing Potential Exemplars in Reducing Zero-Dose Children: A Novel Approach for Identifying Positive Outliers in Decreasing National Levels and Geographic Inequalities in Unvaccinated Children. Vaccines (Basel). 2023 Mar 1;11(3). doi:10.3390/vaccines11030647 Johri M, Rajpal S, Kim R, Subramanian S V. Small-area variation in child under-vaccination in India: a multilevel analysis of cross-sectional data from 36 states and Union Territories, 707 districts, and 22,349 small-area clusters. The Lancet Regional Health - Southeast Asia. 2025 Jan 1;32. doi:10.1016/j.lansea.2024.100504 Tamir TT, Zegeye AF, Mekonen EG, Tekeba B, Ali MS, Gonete AT, et al. Prevalence, spatial variation and determinants of zero-dose children in Ethiopia: Spatial and multilevel analyses. Public Health. 2024 Nov 1;236:365–72. doi:10.1016/j.puhe.2024.09.011 PubMed PMID: 39303624. Utazi CE, Aheto JMK, Chan HMT, Tatem AJ, Sahu SK. Conditional probability and ratio-based approaches for mapping the coverage of multi-dose vaccines. Stat Med. 2022 Dec 20;41(29):5662–78. doi:10.1002/sim.9586 PubMed PMID: 36129171. Agimas MC, Asmamaw M, Hailu MK, Kidie T, Abuhay HW, Yismaw GA, et al. Geospatial mapping to assess the distribution and determinants of zero dose vaccination status hot spots among children in Ethiopia using EDHS 2019: Spatial and geographical weighted regression. PLoS One. 2024 Dec 1;19(12). doi:10.1371/journal.pone.0312610 PubMed PMID: 39739825. Ikilezi G, Augusto OJ, Sbarra A, Sherr K, Dieleman JL, Lim SS. Determinants of geographical inequalities for DTP3 vaccine coverage in sub-Saharan Africa. Vaccine. 2020 Apr 16;38(18):3447–54. doi:10.1016/j.vaccine.2020.03.005 PubMed PMID: 32204938. Nilsen K, Tejedor-Garavito N, Leasure DR, Utazi CE, Ruktanonchai CW, Wigley AS, et al. A review of geospatial methods for population estimation and their use in constructing reproductive, maternal, newborn, child and adolescent health service indicators. BMC Health Serv Res. 2021 Sep 1;21. doi:10.1186/s12913-021-06370-y PubMed PMID: 34511089. Ilesanmi MM, Adeyinka DA, Olakunde BO. Sustainable Development Goals and childhood measles vaccination in Ekiti State, Nigeria: Results from spatial and interrupted time series analyses. Vaccine. 2022 Jun 21;40(28):3861–8. doi:10.1016/j.vaccine.2022.05.037 PubMed PMID: 35644673. Singh R, Agarwal S, De Figueiredo A, Mishra M, Agarwal D. Quantifying the role of maternal recall in estimates of routine immunisation rates in India: a large-scale sub-national Bayesian modelling study. In: COMPASS 2024 - Proceedings of the ACM SIGCAS/SIGCHI Conference on Computing and Sustainable Societies. Association for Computing Machinery, Inc; 2024. p. 74–84. doi:10.1145/3674829.3675062 Gamachu M, Mussa I, Deressa A, Tolera M, Birhanu A, Getachew T, et al. Patterns of basic pneumococcal conjugated vaccine coverage in Ethiopia from 2015 to 2018; further analysis of Ethiopian DHS (2016–2019). Vaccine: X. Elsevier Ltd; 2024. doi:10.1016/j.jvacx.2024.100428 Tesfa GA, Yehualashet DE, Getnet A, Bimer KB, Seboka BT. Spatial distribution of complete basic childhood vaccination and associated factors among children aged 12–23 months in Ethiopia. A spatial and multilevel analysis. PLoS One. 2023 Jan 1;18(1 January). doi:10.1371/journal.pone.0279399 PubMed PMID: 36608032. Artur Manuel Muloliwa, Luiz Antonio Bastos Camacho, José Fernando Souza Verani, Taynãna César Simões, Martinho do Carmo Dgedge. Impact of vaccination on the incidence of measles in Mozambique in the period 2000 to 2011. Cad Saude Publica. 2013. doi:10.1590/s0102-311x2013000200013 Agimas MC, Asmamaw M, Mekonen N, Mamo F, Shewaye DA. Spatial pattern and associated factors of timely vaccination in Ethiopia using EDHS-2016 data: A multilevel and spatial analysis. PLoS One. 2024 Feb 1;19(2 February). doi:10.1371/journal.pone.0296123 PubMed PMID: 38324535. 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 Sep 5;20(1). doi:10.1186/s12889-020-09461-3 PubMed PMID: 32891120. Awol M, Alemu ZA, Moges NA, Jemal K. Geographical variations and associated factors of defaulting from immunization among children aged 12 to 23 months in Ethiopia: using spatial and multilevel analysis of 2016 Ethiopian Demographic and Health Survey. Environ Health Prev Med. 2021 Dec 1;26(1). doi:10.1186/s12199-021-00984-8 PubMed PMID: 34118886. Bantie GM, Tadege M, Nigussie TZ, Woya AA, Tekile AK, Melese AA, et al. Regional disparities of full pentavalent vaccine uptake and the determinants in Ethiopia: Mapping and spatial analysis using the EDHS data. PLoS One. 2025 Jan 1;20(1 January). doi:10.1371/journal.pone.0312514 PubMed PMID: 39787194. 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 Vaccin Immunother. 2022;18(1). doi:10.1080/21645515.2022.2035558 PubMed PMID: 35148252. Shiferie F, Gebremedhin S, Andargie G, Tsegaye DA, Alemayehu WA, Fenta TG. Low Measles Vaccination Coverage and Spatial Analysis of High Measles Vaccination Dropout in Ethiopia’s Underprivileged Areas. Vaccines (Basel). 2024 Mar 1;12(3). doi:10.3390/vaccines12030328 Agegnehu CD, Alem AZ. Exploring spatial variation in BCG vaccination among children 0-35 months in Ethiopia: Spatial analysis of Ethiopian Demographic and Health Survey 2016. BMJ Open. BMJ Publishing Group; 2021. doi:10.1136/bmjopen-2020-043565 PubMed PMID: 33910946. Hailegebireal AH, Hailegebreal S, Tirore LL, Wolde BB. Spatial variation and predictors of incomplete pneumococcal conjugate vaccine (PCV) uptake among children aged 12–35 months in Ethiopia: spatial and multilevel analyses. Front Public Health. 2024;12. doi:10.3389/fpubh.2024.1344089 PubMed PMID: 38864011. Akinyemi YC. Geospatial Analysis of Vaccination Coverage and Predictors of Non- and Under-Vaccination of Children Aged 12–23 Months in Nigeria. Papers in Applied Geography. 2021;7(1):1–33. doi:10.1080/23754931.2020.1809009 Khan J, Shil A, Prakash R. Exploring the spatial heterogeneity in different doses of vaccination coverage in India. PLoS One. 2018 Nov 1;13(11). doi:10.1371/journal.pone.0207209 PubMed PMID: 30485291. Dhalaria P, Kumar P, Verma A, Priyadarshini P, Kumar Singh A, Tripathi B, et al. Exploring landscape of measles vaccination coverage: A step towards measles elimination goal in India. Vaccine. 2024 Jun 20;42(17):3637–46. doi:10.1016/j.vaccine.2024.04.075 PubMed PMID: 38704248. Tamir TT, Terefe B, Wassie M, Workneh BS, Zegeye AF. Prevalence, spatial distribution and determinants of complete childhood pneumococcal conjugate vaccination in Ethiopia: spatial and multilevel analyses. BMC Public Health. 2024 Dec 1;24(1). doi:10.1186/s12889-024-19926-4 PubMed PMID: 39243021. Kamal A, Waseem A, Siddiqa M, Ijaz M, Shakeel A, Iftikhar S. Contextual factors influencing incomplete immunization and investigation of its geospatial heterogeneity in Pakistan: a cross-sectional study based on PDHS (2017–18). BMC Public Health. 2023 Dec 1;23(1). doi:10.1186/s12889-023-16508-8 PubMed PMID: 37620868. Forzy T, Tesfaye L, Getnet F, Misganew A, Lamma SW, Worku A, et al. Assessing the geographic and socioeconomic determinants of vaccine coverage in Ethiopia: A spatial and multistage analysis at the district level. Vaccine. 2025 Apr 19;53. doi:10.1016/j.vaccine.2025.126834 PubMed PMID: 40056895. Diress F, Negesse Y, Worede DT, Bekele Ketema D, Geitaneh W, Temesgen H. Multilevel and geographically weighted regression analysis of factors associated with full immunization among children aged 12–23 months in Ethiopia. Sci Rep. 2024 Dec 1;14(1). doi:10.1038/s41598-024-74189-8 PubMed PMID: 39349770. Tamir TT, Workneh BS, Mekonen EG, Zegeye AF. Coverage, spatial distribution and determinants of childhood inactivated poliovirus vaccine immunization in Ethiopia. PLoS One. 2024 May 1;19(5 May). doi:10.1371/journal.pone.0301933 PubMed PMID: 38820454. Zegeye AF, Mekonen EG, Tekeba B, Alemu TG, Ali MS, Gonete AT, et al. Spatial distribution and determinants of measles vaccination dropout among under-five children in Ethiopia: A spatial and multilevel analysis of 2019 Ethiopian demographic and health survey. PLoS One. 2024 Jul 1;19(7 July). doi:10.1371/journal.pone.0305393 PubMed PMID: 38976660. Defar A, Okwaraji YB, Tigabu Z, Persson LÅ, Alemu K. Geographic differences in maternal and child health care utilization in four Ethiopian regions; A cross-sectional study. Int J Equity Health. 2019 Nov 12;18(1). doi:10.1186/s12939-019-1079-y PubMed PMID: 31718658. Agimas MC, Belew AK, Sisay M, Baffa LD, Gashaw M, Abriham ZY, et al. Spatial variations and determinants of timely completion of vaccination in Ethiopia using further analysis of EDHS 2019 data: Spatial and multilevel analysis. PLoS One. 2024 Apr 1;19(4 April). doi:10.1371/journal.pone.0301409 PubMed PMID: 38578785. Root ED, Lucero M, Nohynek H, Anthamatten P, Thomas DSK, Tallo V, et al. Distance to health services affects local-level vaccine efficacy for pneumococcal conjugate vaccine (PCV) among rural Filipino children. Proc Natl Acad Sci U S A. 2014 Mar 4;111(9):3520–5. doi:10.1073/pnas.1313748111 PubMed PMID: 24550454. Puri P, Khan J, Shil A, Ali M. A cross-sectional study on selected child health outcomes in India: Quantifying the spatial variations and identification of the parental risk factors. Sci Rep. 2020 Dec 1;10(1). doi:10.1038/s41598-020-63210-5 PubMed PMID: 32313048. Teshale AB, Amare T. Exploring spatial variations and the individual and contextual factors of uptake of measles-containing second dose vaccine among children aged 24 to 35 months in Ethiopia. PLoS One. 2023 Jan 1;18(1 January). doi:10.1371/journal.pone.0280083 PubMed PMID: 36598928. Khan J, Shil A, Puri P. Population and sub-national (district) level diversity in missed and dropout of different doses of hepatitis-B vaccine among Indian children aged 12–59 months. PLOS Global Public Health. 2022 May 17;2(5). doi:10.1371/journal.pgph.0000243 Kumar G, R.S. R. Availability of public health facilities and utilization of maternal and child health services in districts of India. Clin Epidemiol Glob Health. 2022 May 1;15. doi:10.1016/j.cegh.2022.101070 Panda BK, Kumar G, Mishra S. Understanding the full-immunization gap in districts of India: A geospatial approach. Clin Epidemiol Glob Health. 2020 Jun 1;8(2):536–43. doi:10.1016/j.cegh.2019.11.010 Tegegne KT, Tegegne ET, Tessema MK, Wudu TK, Chalachew AA, Workneh AZ. Spatial Distribution and Associated Factors of Measles Vaccination Among Children Aged 12–23 Months in Mozambique 2022–2023: A Spatial and Multilevel Analysis. Bratislava Medical Journal. 2025 Feb 1;126(2):181–95. doi:10.1007/s44411-024-00008-1 Das T. Unveiling Spatial Heterogeneity: A Study of Diverse Child Immunization Indicators Across Indian Districts. Child Indic Res. 2024 Aug 1;17(4):1615–59. doi:10.1007/s12187-024-10135-4 Singh SK, Vishwakarma D. Spatial heterogeneity in the coverage of full immunization among children in India: Exploring the contribution of immunization card. Child Youth Serv Rev. 2021 Feb 1;121. doi:10.1016/j.childyouth.2020.105701 Geremew TT, Gezie LD, Abejie AN. Geographical variation and associated factors of childhood measles vaccination in Ethiopia: A spatial and multilevel analysis. BMC Public Health. 2019 Aug 30;19(1). doi:10.1186/s12889-019-7529-z PubMed PMID: 31470822. Chauhadry IA, Soofi SB, Sajid M, Ali R, Khan A, Naqvi SK, et al. Bridging the Vaccination Equity Gap: A Community-Driven Approach to Reduce Vaccine Inequities in Polio High-Risk Areas of Pakistan. Vaccines (Basel). 2024 Dec 1;12(12). doi:10.3390/vaccines12121340 Khan J, Shil A, Mohanty SK. Hepatitis B vaccination coverage across India: Exploring the spatial heterogeneity and contextual determinants. BMC Public Health. 2019 Sep 12;19(1). doi:10.1186/s12889-019-7534-2 PubMed PMID: 31510967. Uthman OA, Adedokun ST, Olukade T, Watson S, Adetokunboh O, Adeniran A, et al. Children who have received no routine polio vaccines in Nigeria: Who are they and where do they live? Hum Vaccin Immunother. 2017 Sep 2;13(9):2111–22. doi:10.1080/21645515.2017.1336590 PubMed PMID: 28665749. Brownwright TK, Dodson ZM, Van Panhuis WG. Spatial clustering of measles vaccination coverage among children in sub-Saharan Africa. BMC Public Health. 2017 Dec 15;17(1). doi:10.1186/s12889-017-4961-9 PubMed PMID: 29246217. Bantie B, Atnafu Gebeyehu N, Adella GA, Ambaw Kassie G, Mengstie MA, Abebe EC, et al. Mapping geographical inequalities of incomplete immunization in Ethiopia: a spatial with multilevel analysis. Front Public Health. 2024;12. doi:10.3389/fpubh.2024.1339539 PubMed PMID: 38912271. Dimitrova A, Carrasco-Escobar G, Richardson R, Benmarhnia T. Essential childhood immunization in 43 low- and middle-income countries: Analysis of spatial trends and socioeconomic inequalities in vaccine coverage. PLoS Med. 2023 Jan 1;20(1). doi:10.1371/journal.pmed.1004166 PubMed PMID: 36649359. Rerolle F, Dey AK, Benmarhnia T, Arnold BF. Spatial targeting and integration across vaccination, vitamin A and deworming programs throughout India 2019–21. Int J Epidemiol. 2024 Dec 1;53(6). doi:10.1093/ije/dyae160 PubMed PMID: 39657980. Umer MF, Zofeen S, Hu W, Qi X, Zhuang G. Spatiotemporal clustering analysis of Expanded Program on Immunization (EPI) vaccination coverage in Pakistan. Sci Rep. 2020 Dec 1;10(1). doi:10.1038/s41598-020-67839-0 PubMed PMID: 32620798. Ketsela A, Gebreyesus SH, Deressa W. Spatial distribution of under immunization among children 12–23 months old in Butajira HDSS, southern Ethiopia. BMC Pediatr. 2021 Dec 1;21(1). doi:10.1186/s12887-021-02690-4 PubMed PMID: 33971837. Utazi CE, Nilsen K, Pannell O, Dotse-Gborgbortsi W, Tatem AJ. District-level estimation of vaccination coverage: Discrete vs continuous spatial models. Stat Med. 2021 Apr 1;40(9):2197–211. doi:10.1002/sim.8897 PubMed PMID: 33540473. Alegana VA, Atkinson PM, Pezzulo C, Sorichetta A, Weiss D, Bird T, et al. Fine resolution mapping of population age-structures for health and development applications. J R Soc Interface. 2015 Apr 6;12(105). doi:10.1098/rsif.2015.0073 PubMed PMID: 25788540. Colson KE, Dwyer-Lindgren L, Achoki T, Fullman N, Schneider M, Mulenga P, et al. Benchmarking health system performance across districts in Zambia: A systematic analysis of levels and trends in key maternal and child health interventions from 1990 to 2010. BMC Med. 2015 Apr 2;13(1). doi:10.1186/s12916-015-0308-5 PubMed PMID: 25889124. Pramanik S, Muthusamy N, Gera R, Laxminarayan R. Vaccination coverage in India: A small area estimation approach. Vaccine. 2015 Mar 30;33(14):1731–8. doi:10.1016/j.vaccine.2015.01.083 PubMed PMID: 25681661. Roberts DA, Ng M, Ikilezi G, Gasasira A, Dwyer-Lindgren L, Fullman N, et al. Benchmarking health system performance across regions in Uganda: A systematic analysis of levels and trends in key maternal and child health interventions, 1990-2011. BMC Med. 2015 Dec 3;13(1). doi:10.1186/s12916-015-0518-x PubMed PMID: 26631048. Wollum A, Burstein R, Fullman N, Dwyer-Lindgren L, Gakidou E. Benchmarking health system performance across states in Nigeria: A systematic analysis of levels and trends in key maternal and child health interventions and outcomes, 2000-2013. BMC Med. 2015 Sep 2;13(1). doi:10.1186/s12916-015-0438-9 PubMed PMID: 26329607. Dong TQ, Wakefield J. SPACE-TIME SMOOTHING MODELS FOR SUBNATIONAL MEASLES ROUTINE IMMUNIZATION COVERAGE ESTIMATION WITH COMPLEX SURVEY DATA. Annals of Applied Statistics. 2021 Dec 1;15(4):1959–79. doi:10.1214/21-AOAS1474 Mercer LD, Safdar RM, Ahmed J, Mahamud A, Khan MM, Gerber S, et al. Spatial model for risk prediction and sub-national prioritization to aid poliovirus eradication in Pakistan. BMC Med. 2017 Oct 11;15(1). doi:10.1186/s12916-017-0941-2 PubMed PMID: 29017491. Sbarra AN, Rolfe S, Nguyen JQ, Earl L, Galles NC, Marks A, et al. Mapping routine measles vaccination in low- and middle-income countries. Nature. 2021 Jan 21;589(7842):415–9. doi:10.1038/s41586-020-03043-4 Atalell KA, Alemayehu MA, Teshager NW, Belay GM, Alemu TG, Anlay DZ, et al. Mapping BCG vaccination coverage in Ethiopia between 2000 and 2019. BMC Infect Dis. 2022 Dec 1;22(1). doi:10.1186/s12879-022-07547-4 PubMed PMID: 35739462. Utazi CE, Thorley J, Alegana VA, Ferrari MJ, Takahashi S, Metcalf CJE, et al. Mapping vaccination coverage to explore the effects of delivery mechanisms and inform vaccination strategies. Nat Commun. 2019 Dec 1;10(1). doi:10.1038/s41467-019-09611-1 PubMed PMID: 30967543. Wariri O, Utazi CE, Okomo U, Metcalf CJE, Sogur M, Fofana S, et al. Mapping the timeliness of routine childhood vaccination in The Gambia: A spatial modelling study. Vaccine. 2023 Sep 7;41(39):5696–705. doi:10.1016/j.vaccine.2023.08.004 PubMed PMID: 37563051. Atalell KA, Liyew AM, Alene KA. Spatial distribution of rotavirus immunization coverage in Ethiopia: a geospatial analysis using the Bayesian approach. BMC Infect Dis. 2022 Dec 1;22(1). doi:10.1186/s12879-022-07825-1 PubMed PMID: 36352357. Sbarra AN, Rolfe S, Haeuser E, Nguyen JQ, Adamu A, Adeyinka D, et al. Estimating vaccine coverage in conflict settings using geospatial methods: a case study in Borno state, Nigeria. Sci Rep. 2023 Dec 1;13(1). doi:10.1038/s41598-023-37947-8 PubMed PMID: 37422502. Wolde HF, Clements ACA, Gilmour B, Alene KA. Spatial co-distribution of tuberculosis prevalence and low BCG vaccination coverage in Ethiopia. Sci Rep. 2024 Dec 1;14(1). doi:10.1038/s41598-024-68549-7 PubMed PMID: 39738221. Utazi CE, Wagai J, Pannell O, Cutts FT, Rhoda DA, Ferrari MJ, et al. Geospatial variation in measles vaccine coverage through routine and campaign strategies in Nigeria: Analysis of recent household surveys. Vaccine. 2020 Mar 23;38(14):3062–71. doi:10.1016/j.vaccine.2020.02.070 PubMed PMID: 32122718. Utazi CE, Olowe ID, Chan HMT, Dotse-Gborgbortsi W, Wagai J, Umar JA, et al. Geospatial Variation in Vaccination Coverage and Zero-Dose Prevalence at the District, Ward and Health Facility Levels Before and After a Measles Vaccination Campaign in Nigeria. Vaccines (Basel). 2024 Dec 1;12(12). doi:10.3390/vaccines12121299 Ferreira LZ, Utazi CE, Huicho L, Nilsen K, Hartwig FP, Tatem AJ, et al. Geographic inequalities in health intervention coverage – mapping the composite coverage index in Peru using geospatial modelling. BMC Public Health. 2022 Dec 1;22(1). doi:10.1186/s12889-022-14371-7 PubMed PMID: 36397019. Dong TQ, Wakefield J. Modeling and presentation of vaccination coverage estimates using data from household surveys. Vaccine. 2021 Apr 28;39(18):2584–94. doi:10.1016/j.vaccine.2021.03.007 PubMed PMID: 33824039. Takahashi S, Metcalf CJE, Ferrari MJ, Tatem AJ, Lessler J. The geography of measles vaccination in the African Great Lakes region. Nat Commun. 2017 May 25;8. doi:10.1038/ncomms15585 PubMed PMID: 28541287. Kawakatsu Y, Mosser JF, Adolph C, Baffoe P, Cheshi F, Aiga H, et al. High-resolution mapping of essential maternal and child health service coverage in Nigeria: a machine learning approach. BMJ Open. 2024 Jun 10;14(6). doi:10.1136/bmjopen-2023-080135 PubMed PMID: 38858137. Utazi CE, Thorley J, Alegana VA, Ferrari MJ, Nilsen K, Takahashi S, et al. A spatial regression model for the disaggregation of areal unit based data to high-resolution grids with application to vaccination coverage mapping. Stat Methods Med Res. 2019 Nov 1;28(10–11):3226–41. doi:10.1177/0962280218797362 PubMed PMID: 30229698. Tesfaye L, Forzy T, Getnet F, Misganaw A, Woldekidan MA, Wolde AA, et al. Estimating immunization coverage at the district level: A case study of measles and diphtheria-pertussis-tetanus-Hib-HepB vaccines in Ethiopia. PLOS Global Public Health. 2024 Jul 25;4(7). doi:10.1371/journal.pgph.0003404 R: The R Project for Statistical Computing [Internet]. [cited 2026 Mar 18]. Available from: https://www.r-project.org/ Statistical software for data science | Stata [Internet]. [cited 2026 Mar 18]. Available from: https://www.stata.com/ Spatial without Compromise · QGIS Web Site [Internet]. [cited 2025 Nov 28]. Available from: https://qgis.org/ Immunization | UNICEF [Internet]. [cited 2025 Nov 24]. Available from: https://www.unicef.org/immunization GAVI. Gavi 2024 Annual Progress Report [Internet]. 2025. Available from: www.gavi.org/funding/financial-reports Homepage | Institute for Health Metrics and Evaluation [Internet]. [cited 2025 Nov 28]. Available from: https://www.healthdata.org/ Nigeria VaxPop II - WorldPop [Internet]. [cited 2025 Nov 28]. Available from: https://www.worldpop.org/current-projects/completed-projects/nigeria-vaxpop-ii/ Ngo-Bebe D, Mechael P, Kwilu FN, Bukele TK, Langwana F, Lobukulu GL, et al. Assessing the use of geospatial data for immunization program implementation and associated effects on coverage and equity in the Democratic Republic of Congo. BMC Public Health. 2025 Dec 1;25(1). doi:10.1186/s12889-025-21578-x PubMed PMID: 39856652. Wonodi C, Farrenkopf BA. Defining the Zero Dose Child: A Comparative Analysis of Two Approaches and Their Impact on Assessing the Zero Dose Burden and Vulnerability Profiles across 82 Low- and Middle-Income Countries. Vaccines (Basel). 2023 Oct 1;11(10). doi:10.3390/vaccines11101543 Hogan D, Gupta A. Why Reaching Zero-Dose Children Holds the Key to Achieving the Sustainable Development Goals. Vaccines (Basel). 2023;11(4). doi:10.3390/vaccines11040781 Wonodi C, Farrenkopf BA. Defining the Zero Dose Child: A Comparative Analysis of Two Approaches and Their Impact on Assessing the Zero Dose Burden and Vulnerability Profiles across 82 Low- and Middle-Income Countries. Vaccines (Basel). 2023 Oct 1;11(10). doi:10.3390/vaccines11101543 The DHS Program - Quality information to plan, monitor and improve population, health, and nutrition programs [Internet]. [cited 2025 Nov 28]. Available from: https://www.dhsprogram.com/ Home - DHIS2 [Internet]. [cited 2025 Nov 28]. Available from: https://dhis2.org/ Rozman LM, Sartori AMC, Souza AA, Soarez PC de. Discrepancy between official records of the Brazilian National Immunization Program Information System and the household survey about vaccination coverage in Cubatão, São Paulo: cross-sectional study, 2023. Epidemiologia e Serviços de Saúde. 2025;34. doi:10.1590/s2237-96222025v34e20240523.en PubMed PMID: 40767724. Arambepola R, Keddie SH, Collins EL, Twohig KA, Amratia P, Bertozzi-Villa A, et al. Spatiotemporal mapping of malaria prevalence in Madagascar using routine surveillance and health survey data. Sci Rep. 2020 Dec 1;10(1). doi:10.1038/s41598-020-75189-0 PubMed PMID: 33093622. Molenaar J, Kikula A, van Olmen J, Pembe A, Beňová L. Getting the numbers right: Power, creativity and ‘good’ routine maternal and neonatal health data in Southern Tanzania. Soc Sci Med. 2025 Feb 1;366. doi:10.1016/j.socscimed.2024.117668 PubMed PMID: 39787826. Bharti N, Djibo A, Tatem AJ, Grenfell BT, Ferrari MJ. Measuring populations to improve vaccination coverage. Sci Rep. 2016;5. doi:10.1038/srep34541 Beaulieu A, Ducharme J, Thibeault C, Akani BC, Ziegler D, Hogan D, et al. The Immunization Agenda 2030 Strategy to reach zero-dose children in low-income and middle-income countries: a scoping review. BMJ Glob Health. 2025 Aug 5;10(8):e018293. doi:10.1136/BMJGH-2024-018293 PubMed PMID: 40763981. Diggle PJ, Tawn JA, Moyeed RA. Model-Based Geostatistics. J R Stat Soc Ser C Appl Stat. 1998 Sep 1;47(3):299–350. doi:10.1111/1467-9876.00113 Giorgi E, Fronterrè C, Macharia PM, Alegana VA, Snow RW, Diggle PJ. Model building and assessment of the impact of covariates for disease prevalence mapping in low-resource settings: to explain and to predict. J R Soc Interface. 2021 Jun 2;18(179):20210104. doi:10.1098/rsif.2021.0104 GAVI. IMPROVING IMMUNISATION COVERAGE AND EQUITY THROUGH THE EFFECTIVE USE OF GEOSPATIAL TECHNOLOGIES Improving Immunisation Coverage and Equity through the Effective Use of Geospatial Technologies and Data [Internet]. 2020 [cited 2025 Nov 28]. Available from: https://www.gavi.org/our-impact/evaluation-studies/using-geospatial-technologies-improve-immunisation-coverage-equity Giorgi E, Macharia PM, Woodmansey J, Snow RW, Rowlingson B. Maplaria: a user friendly web-application for spatio-temporal malaria prevalence mapping. Malaria Journal 2021 20:1. 2021 Dec 20;20(1):471-. doi:10.1186/S12936-021-04011-7 PubMed PMID: 34930265. Johnson O, Fronterre C, Diggle PJ, Amoah B, Giorgi E. MBGapp: A Shiny application for teaching model-based geostatistics to population health scientists. PLoS One. 2021 Dec 1;16(12):e0262145. doi:10.1371/JOURNAL.PONE.0262145 PubMed PMID: 34972193. Wu Y, Dong Q, Xu J, Li ZR, Wakefield J. sae4health: An R Shiny Application for Small Area Estimation in Low- and Middle-Income Countries. ArXiv [Internet]. 2025 May 2 [cited 2025 Nov 28]. Available from: https://arxiv.org/pdf/2505.01467 Espinosa L, Salathé M. Use of large language models as a scalable approach to understanding public health discourse. PLOS Digital Health. 2024 Oct 1;3(10):e0000631. doi:10.1371/JOURNAL.PDIG.0000631 Burstein R, Mafuta E, Proctor JL. Large language models for analyzing open text in global health surveys: why children are not accessing vaccine services in the Democratic Republic of the Congo. Int Health. 2025 Sep 1;17(5):843–52. doi:10.1093/INTHEALTH/IHAF015 PubMed PMID: 40052518. Additional Declarations The authors declare no competing interests. Supplementary Files Supplementaryfile1.pdf Supplementary file 1 Supplementaryfile2.pdf Supplementary file 2 Supplementaryfile3.pdf Supplementary file 3 PRISMAScRChecklist.pdf PRISMA-ScR-Checklist 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-9304718","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Systematic Review","associatedPublications":[],"authors":[{"id":616699782,"identity":"c7eaf232-75b3-4896-a190-b69361253dcd","order_by":0,"name":"Ann M. Njogu","email":"","orcid":"","institution":"Department of Public Health, Institute of Tropical Medicine (ITM), Antwerp, Belgium","correspondingAuthor":false,"prefix":"","firstName":"Ann","middleName":"M.","lastName":"Njogu","suffix":""},{"id":616715560,"identity":"4816fc80-8c54-4200-8f47-232f0ef6bb0c","order_by":1,"name":"Moses M. Musau","email":"","orcid":"","institution":"Population and Health Impact Surveillance Group, Kenya Medical Research Institute-Wellcome Trust Research Programme, Nairobi, Kenya","correspondingAuthor":false,"prefix":"","firstName":"Moses","middleName":"M.","lastName":"Musau","suffix":""},{"id":616715561,"identity":"bf84b229-a2b7-452d-a383-68ddb6551cac","order_by":2,"name":"Swati Srivastava","email":"","orcid":"","institution":"Heidelberg Institute of Global Health, Medical Faculty and University Hospital, Heidelberg University, Heidelberg, Germany","correspondingAuthor":false,"prefix":"","firstName":"Swati","middleName":"","lastName":"Srivastava","suffix":""},{"id":616715562,"identity":"0b5542f1-2994-4c67-a939-a0e6d9735b4e","order_by":3,"name":"Francesco Menegale","email":"","orcid":"","institution":"Center for Health Emergencies, Fondazione Bruno Kessler, Trento, Italy","correspondingAuthor":false,"prefix":"","firstName":"Francesco","middleName":"","lastName":"Menegale","suffix":""},{"id":616715563,"identity":"b1cabffc-92ae-4445-b9ad-8ebb443426dd","order_by":4,"name":"Emma Clarke","email":"","orcid":"","institution":"Swiss Tropical \u0026 Public Health Institute, Basel, Switzerland","correspondingAuthor":false,"prefix":"","firstName":"Emma","middleName":"","lastName":"Clarke","suffix":""},{"id":616715564,"identity":"be66db79-4f2d-4f5f-a071-bf5060f09399","order_by":5,"name":"Caroline Mudereri","email":"","orcid":"","institution":"Centre for Impact, Innovation and Capacity Building for Health Information Systems and Nutrition (CIICHIN), Kigali, Rwanda","correspondingAuthor":false,"prefix":"","firstName":"Caroline","middleName":"","lastName":"Mudereri","suffix":""},{"id":616715565,"identity":"c06f6c91-1457-465c-a6d6-93d1a9244ae7","order_by":6,"name":"Felix R. Kitema","email":"","orcid":"","institution":"Centre for Impact, Innovation and Capacity Building for Health Information Systems and Nutrition (CIICHIN), Kigali, Rwanda","correspondingAuthor":false,"prefix":"","firstName":"Felix","middleName":"R.","lastName":"Kitema","suffix":""},{"id":616715566,"identity":"55f89297-85ff-48bd-8e70-642ce67474b4","order_by":7,"name":"Yvonne Opanga","email":"","orcid":"","institution":"9.\tAmref Health Africa, Kenya","correspondingAuthor":false,"prefix":"","firstName":"Yvonne","middleName":"","lastName":"Opanga","suffix":""},{"id":616715567,"identity":"435474f7-d3ed-44e2-80c5-53fa259d35fb","order_by":8,"name":"Arnold Mmbando","email":"","orcid":"","institution":"Centre for Impact, Innovation and Capacity Building for Health Information Systems and Nutrition (CIICHIN), Kigali, Rwanda","correspondingAuthor":false,"prefix":"","firstName":"Arnold","middleName":"","lastName":"Mmbando","suffix":""},{"id":616715568,"identity":"0fc38d88-6476-4e94-90ea-3165f9d463ca","order_by":9,"name":"Absolomon Gashaija","email":"","orcid":"","institution":"Centre for Impact, Innovation and Capacity Building for Health Information Systems and Nutrition (CIICHIN), Kigali, Rwanda","correspondingAuthor":false,"prefix":"","firstName":"Absolomon","middleName":"","lastName":"Gashaija","suffix":""},{"id":616715569,"identity":"9e45b6d8-9449-4e1c-8f39-a46c1bb20522","order_by":10,"name":"Hassan Sibomana","email":"","orcid":"","institution":"Rwanda Biomedical Center, Kigali, Rwanda","correspondingAuthor":false,"prefix":"","firstName":"Hassan","middleName":"","lastName":"Sibomana","suffix":""},{"id":616715570,"identity":"8c566ccf-2dfa-4555-860e-c583d247f9a4","order_by":11,"name":"Jeanine Condo","email":"","orcid":"","institution":"Centre for Impact, Innovation and Capacity Building for Health Information Systems and Nutrition (CIICHIN), Kigali, Rwanda","correspondingAuthor":false,"prefix":"","firstName":"Jeanine","middleName":"","lastName":"Condo","suffix":""},{"id":616715571,"identity":"b7a0f897-6c79-4338-b72e-90b82f7a2663","order_by":12,"name":"Frank Osei","email":"","orcid":"","institution":"Faculty of Geoinformation Science and Earth Observation (ITC), University of Twente","correspondingAuthor":false,"prefix":"","firstName":"Frank","middleName":"","lastName":"Osei","suffix":""},{"id":616715572,"identity":"6d55f8e8-8675-49a4-9771-3e9491e41485","order_by":13,"name":"Lenka Benova","email":"","orcid":"","institution":"Department of Public Health, Institute of Tropical Medicine (ITM), Antwerp, Belgium","correspondingAuthor":false,"prefix":"","firstName":"Lenka","middleName":"","lastName":"Benova","suffix":""},{"id":616715573,"identity":"3bed583f-849e-4689-a4f6-22f58a5503a8","order_by":14,"name":"Justine I. Blanford","email":"","orcid":"","institution":"Faculty of Geoinformation Science and Earth Observation (ITC), University of Twente","correspondingAuthor":false,"prefix":"","firstName":"Justine","middleName":"I.","lastName":"Blanford","suffix":""},{"id":616715574,"identity":"39c2e4bb-149f-4474-83f3-0e1e39b9701c","order_by":15,"name":"Piero Poletti","email":"","orcid":"","institution":"Center for Health Emergencies, Fondazione Bruno Kessler, Trento, Italy","correspondingAuthor":false,"prefix":"","firstName":"Piero","middleName":"","lastName":"Poletti","suffix":""},{"id":616715575,"identity":"2fb5f53e-341f-4596-8b09-38a454a71f31","order_by":16,"name":"Aleksandra Torbica","email":"","orcid":"","institution":"Department of Social and Political Sciences, Bocconi University, Milan, Italy","correspondingAuthor":false,"prefix":"","firstName":"Aleksandra","middleName":"","lastName":"Torbica","suffix":""},{"id":616715576,"identity":"00d3dc7d-14cc-41d1-aed8-0885cfcb4921","order_by":17,"name":"Alessia Melegaro","email":"","orcid":"","institution":"Department of Social and Political Sciences, Bocconi University, Milan, Italy","correspondingAuthor":false,"prefix":"","firstName":"Alessia","middleName":"","lastName":"Melegaro","suffix":""},{"id":616715577,"identity":"1ece7b2f-6433-4ce7-97e8-5e326cc48e4e","order_by":18,"name":"Carlo Federici","email":"","orcid":"","institution":"CERGAS, SDA Bocconi School of Management, Bocconi University, Milan, Italy","correspondingAuthor":false,"prefix":"","firstName":"Carlo","middleName":"","lastName":"Federici","suffix":""},{"id":616715578,"identity":"9b874faa-2055-4af5-a5b3-2dbe1fa56952","order_by":19,"name":"Peter M Macharia","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyklEQVRIiWNgGAWjYJACZiDmYWBvbADSB0jRwnMQpIcELQwMEgmMxGkxFzv88HNBzT0Z/pmP2x9+YbgjR1CL5ew0Y+kZx4p5JG4nNjbLMDwzJqjF4HaCGTMPWwIPA0iLBMPhxAbCWtK/MfP8S+CRv3kQrKWeCC05Zsy8bQk8BjcYGxs/MBxOIMIvOcXSvH0JPIZnEhtnMxg8MyRoi7l0+sbPPN8S7OWOH3/w8UfFHXmCthggc5h5DHCpw6WF8QcROkbBKBgFo2DkAQBFhj7FIPgtfgAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0003-3410-1881","institution":"Department of Public Health, Institute of Tropical Medicine (ITM), Antwerp, Belgium","correspondingAuthor":true,"prefix":"","firstName":"Peter","middleName":"M","lastName":"Macharia","suffix":""}],"badges":[],"createdAt":"2026-04-02 14:56:53","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-9304718/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9304718/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106094924,"identity":"8a855f85-7a75-477b-a1dc-812cff9a68f3","added_by":"auto","created_at":"2026-04-03 11:43:38","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":80116,"visible":true,"origin":"","legend":"\u003cp\u003eThe PRISMA flow diagram showing the different phases of the scoping review from literature search, article screening and exclusion.\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-9304718/v1/564bd7f9dfd398cca5fad7dc.png"},{"id":106094474,"identity":"df4bda61-d41a-4c10-b2f4-eb2fc0147aba","added_by":"auto","created_at":"2026-04-03 11:42:41","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":124846,"visible":true,"origin":"","legend":"\u003cp\u003eNumber of included studies by country.\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-9304718/v1/ea10d8e59f940cc34e4513c9.png"},{"id":106072131,"identity":"a434584f-7c66-4758-9d11-8cb4957c2fea","added_by":"auto","created_at":"2026-04-03 06:45:47","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":24504,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eDistribution of studies by year of publication and information provided: a) both geospatial data and methodological approaches for modelling ZD children and b) only geospatial data on factors associated with immunization coverage. Not all publications from 2025 were included in the review as the search was conducted in April 2025.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-9304718/v1/cdfb1c94885ffd96ecec4b0d.png"},{"id":106072136,"identity":"6fafdf57-9e4e-48f1-88c8-2f2ddc335615","added_by":"auto","created_at":"2026-04-03 06:45:47","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":94604,"visible":true,"origin":"","legend":"\u003cp\u003eGraphical representation of the distribution of categories and sub-categories of data and covariates. The size of each column represents the number of the covariates in the category or sub-category, for instance, most covariates were classified as demographic while few were infrastructural.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-9304718/v1/469aa5bfe00a3fcba898f380.png"},{"id":106095882,"identity":"ee3291ac-d458-4b42-ad39-9e308615c8e4","added_by":"auto","created_at":"2026-04-03 11:51:28","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1189619,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9304718/v1/9e6f37d2-4c2d-41b9-aa47-6c3e1d6bc083.pdf"},{"id":106072129,"identity":"f3a42125-4d3b-4b1c-b378-89321d5a0f2b","added_by":"auto","created_at":"2026-04-03 06:45:47","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1118604,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary file 1\u003c/p\u003e","description":"","filename":"Supplementaryfile1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9304718/v1/107c74184c943773d8965374.pdf"},{"id":106072133,"identity":"735215fc-f995-461a-9ea0-6e3f9023643f","added_by":"auto","created_at":"2026-04-03 06:45:47","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":499438,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary file 2\u003c/p\u003e","description":"","filename":"Supplementaryfile2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9304718/v1/80c74d295d47f10bc9d7a45b.pdf"},{"id":106072132,"identity":"0a7f7281-71af-473c-a35e-737269456fce","added_by":"auto","created_at":"2026-04-03 06:45:47","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":383786,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary file 3\u003c/p\u003e","description":"","filename":"Supplementaryfile3.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9304718/v1/a616016aab8405a5e5814644.pdf"},{"id":106094971,"identity":"2e17d939-f4ff-455b-b302-9549504f1334","added_by":"auto","created_at":"2026-04-03 11:43:49","extension":"pdf","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":293739,"visible":true,"origin":"","legend":"\u003cp\u003ePRISMA-ScR-Checklist\u003c/p\u003e","description":"","filename":"PRISMAScRChecklist.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9304718/v1/1f2f485362f685390311d34d.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eGeospatial approaches for mapping zero-dose children in low- and lower-1 middle-income countries: A scoping review\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003eChildhood immunization is a pillar of public health, significantly reducing morbidity and mortality from vaccine-preventable diseases (VPDs) (1). Since the establishment of the World Health Organization\u0026lsquo;s (WHO) Expanded Programme on Immunization (EPI) in the 1970s, there has been enormous progress in expanding access to life-saving vaccines worldwide, with 85% of children worldwide now receiving the third dose of diphtheria, tetanus, and pertussis (DTP3) containing vaccine, a common measure of immunization system performance (2).\u0026nbsp;However, even in countries with high immunization coverage levels, there are still pockets of zero-dose (ZD) children - those who do not any receive routine immunization (3,4). Operationally, ZD children are defined as those missing the first dose of DTP (DTP1) vaccine by the end of their first year of life (4). This definition is used because DTP1 is universally used in routine immunization, recommended at six weeks, and is thus a reliable indicator of initial contact with routine immunization services (5).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eZD children are consistently under-served across multiple dimensions of the health system. The absence of DTP1 is not only an indicator of missed immunization but also serves as a practical proxy for limited or absent contact with routine health services more broadly (4). As such, ZD status often reflects wider patterns of exclusion, including reduced access to primary healthcare, maternal and child health services, and other essential interventions. The multiple challenges that ZD children face, for example, nutrition, education, water and sanitation, and other essential resources, exacerbate their vulnerability and risk to disease(6).\u003c/p\u003e\n\u003cp\u003eZD children remain vulnerable to VPDs, which can contribute to sustained disease transmission, even in populations with high vaccination coverage (7). By 2021, approximately twenty-five million children worldwide were not fully vaccinated, with 73% classified as ZD (8). Over 80% of these ZD children were in low- and lower-middle-income countries (LLMICs), with sub-Saharan Africa (SSA) bearing a disproportionate burden. India, Nigeria, Ethiopia, the Philippines and the Democratic Republic of Congo (DRC) accounted for 15%, 12%, 6%, 6%, and 3.8% of the global ZD population, respectively (9,10). Large pockets of ZD children in LLMICs are located especially among marginalized communities (11). For example, in 2024, 31% of children under five years in Nigeria were ZD, clustered in the northwestern parts especially in Zamfara, Sokoto and Kebbi states, and in the poorer and less educated subgroups (12). Years of suboptimal vaccine coverage have resulted in an accumulation of under-immunized populations, fuelling outbreaks of VPDs such as measles, whose global cases reached 9.8 million in 2022, leading to about 136,000 deaths (13).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe COVID-19 pandemic accelerated global vaccination challenges, leading to declines in immunization rates. In India, for instance, the coverage of DTP3 dropped from 91% to 85% during the COVID-19 pandemic (14). This reversal temporarily erased over ten years of progress in vaccination, primarily due to pandemic-related disruptions such as lockdowns and the redirection of healthcare resources towards combating COVID-19 (15). Although the 10.9% rise in ZD children between 2019 and 2024 (16) is partially attributed to COVID-19 disruptions, other factors like poverty, inadequate funding for immunization programs, geographic inaccessibility of immunizing health facilities, armed conflicts, weak health infrastructure, vaccine hesitancy, and cultural or religious opposition to vaccination have been cited as long-standing barriers to reaching ZD populations in LLMICs (17,18).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eConsequently, global initiatives such as the Immunization Agenda 2030 (IA2030) (19) and the GAVI Strategy 5.0 aim to achieve immunization equity, with a particular focus on reaching ZD communities (20,21). The IA2030\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eImpact Goal Indicator 2.1 aims at reducing the number of ZD children by 50% by 2030 using 2019 as the baseline year, and targets to maintain a DTP1 coverage of 99% in countries where DTP1 coverage is already at that level (22).\u0026nbsp;To achieve United Nations\u0026rsquo; Sustainable Development Goal (SDG) 3: good health and well-being (23), it is therefore crucial to identify, target, and reach ZD children with equitable and cost-effective interventions to efficiently address immunization disparities (24\u0026ndash;26). Gavi\u0026rsquo;s Identify-Reach-Monitor-Measure-Advocate (IRMMA) framework emphasizes that the first step in reaching ZD children is accurately identifying and characterizing the communities to which they belong (27).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo identify and characterize these communities at high spatial resolution, spatial, statistical, or a combination of both approaches, have been employed (9,11,26). These approaches often integrate covariates (factors, determinants or variables associated with ZD prevalence or low immunization coverage) with vaccination coverage, to generate high-resolution estimates of ZD prevalence. These spatial and spatial-statistical techniques emerged as more effective tools for unmasking disparities in ZD prevalence that are often obscured by regional or national estimates (28,29). In the last two decades, spatial and spatial-statistical approaches used for mapping health outcomes have considerably advanced (30), including for vaccination coverage (31\u0026ndash;33). This has been fuelled by the increasing availability of immunization data from routine health information systems (e.g., District Health Information Software-2 (DHIS2) and household surveys (e.g., Demographic and Health Surveys (DHS) and Multiple Indicator Cluster Surveys (MICS)) that incorporate global positioning systems (GPS) (34,35), as well as geocoded databases of health facilities (36), satellite derived variables and development of advanced methods to synthesize these datasets (30). For example, the number of countries using DHIS2 for health reporting increased from 11 to 99 between 2010 and 2025, while DHS surveys have increased exponentially between 1985 and 2025, now covering 92 out of 129 LMICs (37\u0026ndash;39). However, there is a conspicuous gap in systematic synthesis of spatial data sources, covariates and modelling frameworks used to identify and map ZD and under-immunized children.\u003c/p\u003e\n\u003cp\u003ePrevious literature reviews have mainly focused on factors or determinants (40\u0026ndash;42), barriers (43,44), and inequalities (45\u0026ndash;47) associated with general childhood vaccination coverage or ZD prevalence. While findings of these reviews highlight disparities in vaccine coverage among vulnerable groups (such as migrants, refugees, and those in informal settlements), they do not have a geospatial lens, leaving a critical gap in high-resolution mapping of the heterogeneity in observed coverage levels and in understanding key spatial determinants. So far, only one narrative review has applied a geospatial lens to examine microplanning, geospatial, and machine learning (ML) approaches for reaching ZD and under-immunised children. While it provides a narrative overview of the strengths and limitations of these approaches, it was restricted to 18 publications and reports in SSA only mainly published in in 2024 and 2025, did not explicitly focus on spatial methods, and did not critically assess the methods, data and data sources used in the studies (48).\u003c/p\u003e\n\u003cp\u003eAdopting a geospatial lens is critical for identifying ZD children in LLMICs, as they are often found in clusters that are masked by national or regional averages (49). Geospatial approaches enable the detection of fine-scale geographic inequities, revealing underserved communities affected by remoteness, conflict, urban informality, or weak health system reach (49,50). By integrating spatial analysis with demographic and health data, policymakers and practitioners can more precisely target resources, design context-specific interventions, and improve the efficiency and equity of immunization strategies aimed at reaching ZD children.\u003c/p\u003e\n\u003cp\u003eTo date, no synthesized evidence has described (i) spatial datasets that have been used to represent determinants of ZD and under-immunization, (ii) spatial and spatial-statistical approaches that have been applied to generate ZD estimates in LLMICs and (iii) gaps limiting policy-relevant use of these approaches. To address these gaps, we conducted a scoping literature review covering all LLMICs and outlined the relevant spatial datasets that have been used to map ZD prevalence. Secondly, we highlighted the methods used in mapping ZD children. Further, we outlined gaps in both geospatial data and methodological frameworks used, and the corresponding recommended strategies to address these gaps in most vulnerable populations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eA scoping review is appropriate because evidence on geospatial methods for identifying ZD children is diverse, cross-disciplinary, and evolving. Research varies in terms of data sources, characteristics, and analytical approaches, making traditional systematic reviews impractical. A scoping review enables mapping of the literature, identifying key concepts, gaps, methodological trends, and how geospatial approaches are applied to study ZD populations.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003ch2\u003eSearch Strategy\u003c/h2\u003e\n\u003cp\u003eWe adhered to Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for scoping reviews (PRISMA-ScR) guidelines (51) and the protocol was registered in Open Science Framework (52). A comprehensive search for peer reviewed articles was conducted in April 2025 in six electronic databases: PubMed, Web of Science, Scopus, Cochrane, Embase, and EBSCOhost-CINAHL. Search terms were designed in accordance with the Population-Context-Concept (PCC) framework (53), where Population encompassed all studies about vaccination coverage; Concept referred to studies using spatial or spatial-statistical modelling methods; and Context to studies in LLMICs (Supplementary file 1 table S1).\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eScreening and selection criteria\u003c/h2\u003e\n\u003cp\u003eArticle screening was conducted in two phases: title and abstract screening and full-text review. Screening was conducted independently by two reviewers (AMN and MMM), to minimize bias in the selection of articles. Any disagreements regarding exclusion or inclusion of articles were resolved by a third reviewer (PMM). We included studies published in English that were conducted in LLMICs and used, or explicitly described the use of, spatial data to map or model ZD children or routine childhood immunization coverage with no temporal limits. We also included studies using geospatial methods to predict, visualize, or analyze coverage of routine childhood vaccinations such as DTP1, measles-containing vaccines (MCV), among others. Studies were excluded if they were conducted outside LLMICs, focused on immunization coverage without specific attention to childhood vaccination, or addressed only clinical aspects of vaccine efficacy or safety without examining coverage. We also excluded literature reviews and studies that examined determinants, barriers, or correlates of immunization uptake without incorporating a spatial component. The inclusion and exclusion criteria are detailed in Supplementary file 1 table S2.\u003c/p\u003e\n\u003cp\u003eAlthough the focus of this study was primarily on ZD children, studies assessing child-routine vaccination coverage through geospatial techniques were also retained since the same data and methods apply. Search terms were first piloted on PubMed (Supplementary Box 1)\u0026nbsp;and then adapted to the other databases. Additional grey literature was retrieved from unpublished reports, conference proceedings, websites, dissertations and snowballing of eligible studies(53).\u003c/p\u003e\n\u003ch2\u003eData Extraction\u003c/h2\u003e\n\u003cp\u003eFrom each study, we extracted: i) bibliographic information, including title, authors, Digital Object Identifier (DOI), year of publication, and study objective, ii) study characteristics such as the objective, geographical setting and study period, iii) vaccine-related details, including vaccine type, source of vaccination data and the definition of ZD where the study focused on ZD, iv) types and sources of covariates, v) modelling approaches, including model type, software used, and whether modelling code was available), vi) results, such as spatial and temporal resolution of estimates and spatial aggregation units, and vii) limitations and recommendations reported by the authors. A summary of the variables we extracted from the articles is presented in Supplementary file 1 table S3.\u0026nbsp;\u003c/p\u003e\n\u003ch2\u003eData Synthesis\u003c/h2\u003e\n\u003cp\u003eData synthesis was performed independently by two reviewers (AMN and PMM) using integrated narrative and thematic approaches to synthesize the extracted data, as these methods are well suited to integrate diverse findings from scoping reviews.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe summarized the studies by their geographical distribution, year of publication, whether they provided spatial modelling methods, spatial data or both, the type of vaccines they assessed, the target age groups used to compute coverage and how ZD was defined. Thematic summarization involved inductively identifying recurring themes related to geospatial data and modelling approaches. Modelling methods were hierarchically organized into bespoke or ad-hoc themes and sub-themes based on the core concepts and relative methodological complexity. Covariates identified from the studies were thematically classified into broad categories (such as demographic, and health system) and specific sub-categories like household characteristics factors. This classification was informed primarily by the contextual usage of the covariates in the studies. Modelled estimates were summarized based on the aggregation units (for example, by administrative region or by pixel), spatial and temporal resolution, while tools, packages, reported limitations and recommendations were presented narratively. In addition, we identified limitations beyond those explicitly reported in the studies and proposed recommendations to address these gaps, with the aim of strengthening future ZD prevalence modelling.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003eArticle Search and Screening\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOf 15,587 initially retrieved articles, including 15,586 sourced from databases and one from citation searching, 7,871 duplicates were removed. The remaining 7,716 articles underwent title and abstract screening, resulting in the exclusion of 7,445 articles that did not meet the eligibility criteria (Figure 1). These 7,445 excluded articles were either literature reviews, studies focused on immunization other than routine childhood vaccination (e.g., COVID-19) or studies that did not include or describe spatial data or methods.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOf the 271 articles selected for full-text review, the full text of one article could not be found. We further excluded 169 articles that did not use geospatial methods (n = 151), did not focus on childhood vaccination (n=7) or vaccination coverage (n=5), or focused on disease modelling without addressing vaccination coverage (n=5). Finally, 102 articles were included in the review (Supplementary File 2 table S1).\u003c/p\u003e\n\u003cp\u003eStudy Characteristics\u003c/p\u003e\n\u003cp\u003eAmong all 76 LLMICs (39), 68 (89.5%) countries had at least one study included in this review (Figure 2). Of the 102 studies, 18 (17.6%) were conducted across multiple countries (Supplementary file 1 table S4), while 84 (82.4%) focused on a single country (Supplementary file 1 table S5). Ethiopia (38%), Nigeria (30%), and India (19%) accounted for the largest portion of studies, together representing 87% of all included studies. Of the 84 single-country studies, 90% covered the entire country, while the remainder focused on specific subnational areas, including states, provinces, districts, counties and cities as outlined in Supplementary file 1 table S5.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe included studies were published between 2010 and 2025, with the majority (70%) published between 2021 and 2024 (Figure 3). Most studies (n=86; 84.3%) provided both geospatial data and methodological approaches for modelling ZD children. The remaining 16 studies (15.7%) provided geospatial data on factors associated with immunization coverage but did not explicitly model vaccination coverage. Instead, these studies primarily examined associations or correlations between covariates and vaccination outcomes.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eVaccination\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eVaccines assessed\u003c/p\u003e\n\u003cp\u003eUnique instances of vaccines assessed across the included studies were: Bacillus Calmette\u0026ndash;Gu\u0026eacute;rin (BCG) (n=41), any dose of DTP (n=67), measles containing vaccines (MCV) (n=69), oral polio vaccine (OPV) (n=40), inactivated polio vaccine (IPV) (n=2), pneumococcal conjugate vaccine (PCV) (n=7), rotavirus vaccine (n=9), and Hepatitis B virus vaccine (HBV) (n=4). MCV1 was the most frequently studied vaccine, appearing in 66.7% (n=68) of the studies, followed by DTP1 (n=55) and DTP3 (n=55). A tabulation of vaccines and the corresponding number of studies is presented in Supplementary file 1 table S6,\u0026nbsp;with some assessing more than one vaccine (n=73).\u003c/p\u003e\n\u003cp\u003eAcross all studies, 24 different age groups were used to report coverage of 21 vaccines (Supplementary file 1 table S20). The number of target age groups used for the same vaccine also varied, ranging from fewer than five for 13 vaccines to more than ten for four vaccines. For instance, MCV1 was reported using 21 target age groups across 67 studies, with the most common being children aged 12\u0026ndash;23 months (n=32). Similarly, coverages for DTP1 and DTP3 were reported using ten and twelve age groups, respectively, with children aged 12\u0026ndash;23 months being the most used age group. See Supplementary file 1 table S20 for a detailed tabulation of the age groups.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eZD definition\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOnly 19.6% of the included studies (n=20) focused on ZD prevalence (Supplementary file 1 table S7). Across these 20 studies, ZD was operationalized inconsistently, with two dominant definitions and four target age-groups, limiting comparability across settings. Specifically, a ZD child was defined as: one who had not received DTP1 in 17 studies (6,10,24,50,54\u0026ndash;65) or one missing a single dose in the routine vaccination schedule in three studies (66\u0026ndash;68). Four of the 20 studies only presented descriptive statistics or conducted simple analyses examining the relationship between ZD determinants (but contributed geospatial data) and ZD status (10,26,62,67) without predicting ZD. Of the 20 studies, one spanned the whole of SSA (55) and two covered all low- and middle-income countries (LMICs) (26,64). The rest were conducted mainly in Nigeria (n=8), Ethiopia (n=6), DRC (n=5), India (n=5) and Pakistan (n=4), with four studies spanning multiple countries. Over 50% of these studies (n=13) were published in 2023 or later.\u003c/p\u003e\n\u003cp\u003eFive age groups were used to define the target population for evaluating ZD children, including \u0026lt;1 year (n=2), 0\u0026ndash;23 months (n=1), 12\u0026ndash;23 months (n=12), 12\u0026ndash;35 months (n=3) and \u0026lt;5 years(n=1). One study did not specify the target age groups used to define ZD. The detailed summary of the target age-groups used is presented in Supplementary file 1 table S7.\u003c/p\u003e\n\u003cp\u003eData characteristics and sources\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eVaccination data\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eSurvey data only\u003c/em\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOf the 102 studies, 85 obtained vaccination status data from survey-based sources. These studies primarily relied on DHS data, either as a standalone source (n = 57) or in combination with other household surveys like Multiple Indicator Cluster Surveys (MICS) (n=13). Other studies (n=12) used field survey data collected after Supplementary Immunization Activities (SIA), clinical trials and community demonstration projects (Supplementary file 1 table S8).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAdditionally, three studies used pre-modelled vaccination estimates from the Institute for Health Metrics and Evaluation (IHME)\u0026rsquo;s local burden of disease project (26,64,69) which were derived using data from multiple household surveys.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eRoutine data only\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eEleven studies conducted in Kenya, Nigeria, Ethiopia, Mozambique, Pakistan, Zambia, and Cameroon used routine data only. These studies relied on data from national health information systems such as DHIS2 and health district surveillance systems (Supplementary file 1 table S8).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eStudies combining survey and routine data\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eOnly five studies combined routine and survey data to analyse vaccine coverage. In Ethiopia, two studies integrated DHIS2 and DHS data. In Zambia, one study triangulated routine Health Management Information System (HMIS) data with data from 20 household surveys, including DHS and MICS. In Pakistan, one study combined health facility records with survey data collected after SIA. In Madagascar, one study used vaccination data from health centre EPI records alongside data from a longitudinal cohort study conducted in a single district.\u003cem\u003e\u0026nbsp;\u003c/em\u003eMore details on the sources of vaccination data are outlined in Supplementary file 1 tables S8 \u0026amp; S9.\u003c/p\u003e\n\u003cp\u003eCovariates\u003c/p\u003e\n\u003cp\u003eTypes of covariates\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe extracted details on covariates (factors, determinants or variables associated with ZD prevalence or low immunization coverage) from 81 of the 102 included studies. Of the remaining 21 studies, five modelled vaccination coverage without including covariates (64,65,70\u0026ndash;72), while 15 did not predict vaccine coverage and thus they did not provide covariates (Supplementary file 1 table S10).\u003c/p\u003e\n\u003cp\u003eThe number and types of covariates used varied substantially across studies, with 15 studies using fewer than six covariates and 41 studies including ten or more. Since individual studies had multiple covariates, the percentages reported in this section are based on the total number (n=931) of covariates (unique instances) extracted and not the number of studies.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThere were six general categories of covariates used (Figure 4). Demographic factors, accounted for nearly half (49.4%) of all the covariates used, while health system factors and physical and environmental factors represented 19.2% and 15.9% of all covariates, respectively. Factors representing hard-to-reach contexts, such as exposure to conflict, and remoteness comprised only 8.1% while geographic region, infrastructural factors, such as access to electricity and distance to transport networks, accounted for just 3.9%. Other less frequently used covariates represented spatial characteristics such as the region where the child resided, the geographical size of the region or the health area. These were classified as \u0026ldquo;others\u0026rdquo; and accounted for 4% of all covariates.\u003c/p\u003e\n\u003cp\u003eIn terms of the number of studies using specific categories of covariates, 95% included demographic covariates, 76% included at least one health system\u0026ndash;related factor, and only 40% included a factor representing marginalization or hard-to-reach populations (Supplementary file 1 table S11).\u0026nbsp;The distribution of covariates used in ZD studies only was largely consistent with that observed across all the 102 studies, with a slightly greater representation of variables capturing hard-to-reach populations\u003cem\u003e\u0026nbsp;\u003c/em\u003e(Supplementary file 1 figure S1).\u003c/p\u003e\n\u003cp\u003eA further breakdown of covariates beyond the six classes (Table 1) shows that the most frequently used variables are those linked to healthcare access and utilization, education and awareness, socio-economic status, characteristics of the child and the household and environmental factors. Health financing and covariates representing hardship areas such as presence of conflicts and slums, migration were featured the least.\u003c/p\u003e\n\u003cp\u003eTable 1. Summary of covariates and data inputs across included studies\u0026nbsp;\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"652\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 95px;\"\u003eBroad category\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eSubcategory\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003eSpecific aspects\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eSource of the geospatial data\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" rowspan=\"10\" style=\"width: 95px;\"\u003eDemographic (n=459, 49.3%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003ePopulation (31)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eDensity (22), subset of the population (4) and other characteristics (5)\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eEmpirically modelled (1), Census (3), NGO, census, HMIS, UN agency (1), Global mapping projects (24), Existing survey (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eUrbanicity (50)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eUrban/rural dichotomy (39), urban continuum (7) and settlements (4)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eGlobal mapping projects (5), UN agency (1), Existing survey (31), European Commission (6), Satellite-based (1), University research unit (1), Primary data collection (1), Electronic Immunization Registry (1), Space agency (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eAge (44)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eAge of the care giver (13) and the child (31)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (37), Primary data collection (6)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eSocio-economic (77)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eHousehold wealth/assets (37), occupation (15), poverty (15), economic environment (7) and bank account (3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eEmpirically modelled (2), Existing survey (59), University research unit (1), Primary data collection (3), UN agency (2), Global mapping projects (8)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eEducation and awareness (91)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eMaternal education (42), literacy (12), Media and internet (23), knowledge of disease (4) and other aspects of education (10)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (73), Global mapping projects (4), Primary data collection (8), Electronic Immunization Registry (1), HMIS (1), Empirically modelled (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eCulture and beliefs (34)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eEthnicity (10), religion (19) and social group (5)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (27), University research unit (3), Primary data collection (3), HMIS (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eChild characteristics (53)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eChild\u0026apos;s gender \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; (28), birth order (17) and birth interval (2) and others (6)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (45), Primary data collection (5), Electronic Immunization Registry (1), HMIS (2)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eHousehold characteristics (54)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eHousehold head (17), size (11), marital status (13), number of children (10), and others (3) \u0026nbsp;\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (46), Primary data collection (7), Empirically modelled (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eFertility and reproduction (14)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eFertility \u0026nbsp; \u0026nbsp;3), parity (6), pregnancy (4) and desire for last child (1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (11), Global mapping projects (1), Primary data collection (2)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eNutrition and hygiene (11)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eWater (3), sanitation (3), breast feeding and diet (5)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (9), Empirically modelled (2)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" rowspan=\"5\" style=\"width: 95px;\"\u003eHealth system (n=179, 19.2%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eDisease \u0026amp; morbidity (12)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eChildhood diarrhoea (1), respiratory infections (1), malnutrition (2), malaria (4), Polio \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; 2 and others (2)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eGlobal mapping projects (6), Existing survey (1), Empirically modelled (2), Primary data collection (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eHealth care utilization (91)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eAntenatal care (30), postnatal care (12), place of delivery (23), mode of delivery (3), skilled birth attendance (6), autonomy in decision making (8) and possession of a health card or document (9)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (83), Primary data collection (3), Global mapping projects (1), Electronic Immunization Registry (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eHealth intervention (20)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eBed nets (3), family planning (2) and other vaccinations (15)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (12), Primary data collection (1), Electronic Immunization Registry (1), Empirically modelled (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eAccess to health care services (46)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eProximity to healthcare (32), availability of healthcare facilities and workers (14)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eGlobal mapping projects (13), Primary data collection (9), Existing survey (2), Empirically modelled (13), HMIS (1), HMIS, UN agency (1), Crowdsourced (1), census (1), Satellite-based (1), UN agency (1), Electronic Immunization Registry (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eHealth financing (10)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eHealth insurance (6) and expenditure (4)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (9), Global mapping projects (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 95px;\"\u003eInfrastructural (n=36, 3.9%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eAccess to infrastructure (36)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eTransport infrastructures ( \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;18), electricity (3) and night-time lights as a proxy (15)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eCrowdsourced (8), Global mapping projects (4), Empirically modelled (3), European Commission (2), Existing survey (2), Satellite-based (12), Global climate database (1), Other global databases (roads, boundaries, water bodies) (2)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" rowspan=\"4\" style=\"width: 95px;\"\u003eHard to reach (n=72, 7.7%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eConflict (11)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eConflict areas and proximity to such areas\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eUniversity research unit (1), Global mapping projects (9)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eRemoteness (56)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eAccess to cities (20) and settlements (16) proximity to cultivated areas (6), protected areas (4) and water bodies (10)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eCrowdsourced (9), Satellite-based (1), Empirically modelled (5), Global mapping projects (24), Space agency (8), European Commission (3), Other global databases (roads, boundaries, water bodies) (4), Regional mapping agency (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eMigration (3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eMigration status (3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eExisting survey (1), Primary data collection (1), HMIS (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eSlums (2)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eSlum areas (2)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eGlobal health organization (1), Existing survey (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" rowspan=\"5\" style=\"width: 95px;\"\u003ePhysical and environmental (n=148, 15.9%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eTopographical (21)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eElevation (16) and slope (5)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eSatellite-based (14), Regional mapping agency (1), National mapping agency (3), Global mapping projects (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eHydrological (41)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003ePrecipitation (21), evapotranspiration (8), aridity (5) and soil/vegetation moisture (7)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eGlobal research partnership (7), University research unit (14), Global climate database (9), Satellite-based (7), Empirically modelled (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eAtmospheric conditions (45)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eTemperature (37), cloud cover percentage (3), frost day frequency (3) and mean vapour pressure (2)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eUniversity research unit (17), Satellite-based (16), Global climate database (7)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eVegetation conditions (21)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eEnhanced vegetation index (13), normalized difference vegetation index (5) and other indices (3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eSatellite-based (20)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eAgricultural (20)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eLivestock density (13), irrigation ( \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; 3) and the growing season and suitability (4)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eGlobal mapping projects (10), Primary data collection (1), UN agency (5), University research unit, UN agency (2), University research unit (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" style=\"width: 95px;\"\u003eOthers (n=37, 4%)\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 113px;\"\u003eGeographical location (37)\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 170px;\"\u003eRepresentation of space through coordinates (1), subnational (23) or national regions (13).\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 274px;\"\u003eSatellite-based (1), Other global databases (roads, boundaries, water bodies) (7), Existing survey (21), HMIS (1), Crowdsourced (1), Global health organization (1), Primary data collection (1), Empirically modelled (1)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eFootnote: Values in parentheses represent the number of studies (n)\u003c/p\u003e\n\u003cp\u003eSources of data on covariates\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCovariate data were drawn from a wide range of sources. 51% out of all extracted covariates were sourced from secondary survey data. Other major sources included, university-led global mapping initiatives (15.9%), institutional datasets from mapping agencies and regional or global organizations (8.5%), satellite-derived products (7.8%), and primary surveys (5.6%). Other sources included national censuses, crowd-sourced data, model-derived outputs, and data from HMIS. There were a few instances (n=69 covariates) in which studies triangulated multiple data sources and overall, data sources were not reported for 39 covariates.\u003c/p\u003e\n\u003cp\u003eAmong the covariates obtained from secondary surveys, DHS was the most common source (88%) while MICS accounted for about 10%. Global mapping projects led by universities or independent research groups provided substantial sources of covariates: Worldpop, University of Southampton (33%) and Malaria Atlas Project, Kids Research Institute Australia and Curtin University (19%), and other similar projects such as IHME\u0026rsquo;s local and global burden of disease, Climate Research unit at University of East Anglia, Gridded Livestock of the World by Food and Agriculture Organization of the United Nations (FAO), Armed Conflict Location \u0026amp; Event Data. Global institutional sources included the European Commission, UN agencies, space agencies, and global partnerships. Additionally, the Moderate Resolution Imaging Spectroradiometer (MODIS) and the National Oceanic and Atmospheric Administration (NOAA) provided most satellite-based datasets. A summary of the sources of covariates is provided in Table 1 and Supplementary file 1 table S12.\u003c/p\u003e\n\u003cp\u003ePre-processing of covariates\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSome covariates were available as gridded surfaces which needed little to no pre-processing before being incorporated into geospatial models. However, some were derived from point data and required pre-processing to transform them into continuous surfaces. These included modelling of travel time/distance to service points using spatial accessibility models, estimating exposure to conflict using kernel density estimation approach, spatial interpolation, georeferencing, data quality, completeness and consistency checks, and subsequent imputations. Other pre-processing steps included harmonizing spatial resolution and coordinate reference systems, resampling or aggregating gridded data, and aligning covariates to consistent administrative boundaries. Some studies also screened covariates for multicollinearity and performed variable selection or dimensionality reduction prior to modelling.\u003c/p\u003e\n\u003cp\u003eModelling vaccination coverage\u003c/p\u003e\n\u003cp\u003eStudies clustered around ten broad themes. Most focused on geospatial mapping of vaccination coverage, encompassing the modelling and visualisation of hotspots, under-immunisation, ZD children, and the determinants of spatial inequity. Vaccine timeliness, dropout, and their geographic drivers formed a related area of inquiry. Access to vaccines was examined through multiple proxies, with geographic accessibility and health system reach receiving particular attention in remote, fragile, and conflict-affected settings. Further themes included health system performance, delivery strategies, and the spatial integration of immunisation within broader health programmes. A further area of work addressed data quality and coverage estimation challenges using advanced spatial-statistical methods and machine learning (Supplementary file 2 table S2).\u003c/p\u003e\n\u003cp\u003eModelling methods\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eStudies modelling vaccination coverage (n=86; 92.2%), including ZD prevalence, used a variety of methods ranging from simple spatial clustering approaches to complex approaches such as Bayesian geostatistical methods and ML. These approaches were either spatial (67.4%) or spatiotemporal (32.6%) and were summarized into five categories: i) spatial autocorrelation and clustering approaches, ii) small area estimation (SAE) methods, iii) spatial interpolation techniques, iv) ML methods and v) other methods such as disaggregation and multi-criteria decision analysis (Table 2). In many instances (n=49), studies applied several methods.\u003c/p\u003e\n\u003cp\u003eSpatial autocorrelation \u0026amp; clustering\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOver half of the methodological studies (n=46, 53.5%) focused on assessing spatial autocorrelation and clustering of immunization status either as standalone analysis (n=15, 32.6%) or in combination with other spatial methods (n=31, 67.4%). Global spatial autocorrelation (using global Moran\u0026rsquo;s I) tests the overall patterns of dispersion and correlation, while local spatial autocorrelation (local Moran\u0026rsquo;s I) identifies local clusters of low or high immunization coverage. Further, hotspot analysis with the Getis-Ord Gi* Statistic is used to determine statistically significant hotspots and cold spots of vaccine coverage, showing clusters of high or low immunization coverage. Kulldorff\u0026rsquo;s spatial scan, on the other hand, employs a moving window across the study area to detect statistically significant clustering of areas with higher or lower vaccination coverage than expected.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTable\u0026nbsp;2. Spatial and spatiotemporal methods used to model or map vaccination coverage\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"642\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 113px;\"\u003eGeneral category\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003eSub-category\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eMethod\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003eReference\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"5\" style=\"width: 113px;\"\u003eSpatial autocorrelation, hot spot analysis and cluster detection (n=46)\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003eGlobal spatial autocorrelation\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eGlobal Moran\u0026rsquo;s I (n=37)\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(59,66,68,71,73\u0026ndash;105)\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" style=\"width: 132px;\"\u003eLocal spatial autocorrelation\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eLocal Moran\u0026rsquo;s I (n=15)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e\u003cspan lang=\"EN-GB\"\u003e(73,75,77,78,82,84\u0026ndash;86,88,96,102,103,106\u0026ndash;109)\u003c/span\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 227px;\"\u003eSubtype not specified (n=2)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(71,100)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 227px;\"\u003eGetis-Ord Gi* statistic (n=21)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(59,66,73,76,80\u0026ndash;83,88,90\u0026ndash;94,97,101,109\u0026ndash;111)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 132px;\"\u003eCluster detection\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eKulldorff\u0026rsquo;s spatial scan (n=21)\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(63,66,68,73,74,76\u0026ndash;78,80,82,83,87,88,91,92,94,97,101,104,109,112,113)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 113px;\"\u003eSmall area estimation (n=11)\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003eSpatial model\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eVarious SAE spatial specifications (n=5)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(65,98,107,114,115)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 132px;\"\u003eSpatiotemporal model\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eVarious spatiotemporal SAE model specifications (n=6)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(116\u0026ndash;121)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"9\" style=\"width: 113px;\"\u003eSpatial Interpolation (n=39)\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd rowspan=\"2\" style=\"width: 132px;\"\u003eOther\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eInverse distance weighting (n=1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(59)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 227px;\"\u003eExact method not stated\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e\u003cspan lang=\"EN-GB\"\u003e(105)\u003c/span\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"5\" style=\"width: 132px;\"\u003eKriging\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eUniversal kriging (n=2)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(82,91)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 227px;\"\u003eEmpirical Bayesian kriging (n=3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(78,82,88)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 227px;\"\u003eSimple ordinary (n=1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(88)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 227px;\"\u003eOrdinary kriging (n=10)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e\u003cspan lang=\"EN-GB\"\u003e(68,73,76,77,82,83,88,92,94,109)\u003c/span\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 227px;\"\u003eType not specified (n=6)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(66,87,89,95,97,101)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 132px;\"\u003eGeostatistical models\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eBayesian model-based geostatistics (MBG) (n=21)\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e\u003cspan lang=\"EN-GB\"\u003e(6,7,24,25,50,54,55,58,61,67,72,122\u0026ndash;132)\u003c/span\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 227px;\"\u003eReused outputs derived from a geostatistical model (n=1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(64)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"5\" style=\"width: 113px;\"\u003eArtificial Intelligence (Machine learning (n=7))\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003eAdditive\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eGeneralized Additive Models (GAMs) (n=5)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e\u003cspan lang=\"EN-GB\"\u003e(55,61,122,133,134)\u003c/span\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 132px;\"\u003eTree-based models\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eGradient Boosting Model (GBM), and Boosted regression trees (BRT), Random forests (RF), Decision Tree (DT), (n=6)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(55,61,69,89,122,134)\u003cbr\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 132px;\"\u003eRegularization- linear models\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eLeast Absolute Shrinkage and Selection Operator (LASSO), Ridge classifier (n=3)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e\u0026nbsp;\u003cbr\u003e(55,61,69)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 132px;\"\u003eNeural Networks\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eMulti-Layer Perceptron (MLP) (n=1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e\u003cspan lang=\"EN-GB\"\u003e(56)\u003c/span\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 132px;\"\u003eInstance-based\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eNearest Neighbour (n=1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(56)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"3\" style=\"width: 113px;\"\u003eOthers (n=4)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 132px;\"\u003eDisaggregation\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eSpatial regression (n=1)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(135)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 132px;\"\u003eMulti-criteria decision analysis\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eVulnerability index (n=2)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(6,60)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 132px;\"\u003e\u0026nbsp;\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 227px;\"\u003eAdjusted numerator (routine) and denominator (population) (n=2)\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 170px;\"\u003e(70,136)\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSmall Area Estimation (SAE)\u003c/p\u003e\n\u003cp\u003eSAE modelling frameworks were used in only 11 studies (12.8%) of all modelling studies. These spatial-statistical approaches were used to generate estimates of vaccination coverage for small geographic areas with limited sample sizes (unstable direct estimates) by borrowing strength from neighbouring areas or through use of auxiliary data such as malnutrition, incidence of VPDs like polio, among others. There was marked variation in model specification across SAE studies, including the use of spatial versus spatiotemporal frameworks. Studies also differed in whether they included covariates or not, how they modelled spatial random effects \u0026ndash; both unstructured or/and structured components (e.g., BYM and Leroux models and definition of neighbourhood matrices), incorporation of temporal effects (e.g. autoregressive models or random walks of first or second order), in the spatiotemporal interaction terms they used (such as Type 1), and in the priors they used (e.g. inverse Gamma, Penalized Complexity). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSpatial Interpolation\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSpatial interpolation techniques ranged from simple inverse distance weighting (n=1) and kriging (n=18), to more complex geostatistical modelling (n=22). Kriging \u0026ndash; a spatial interpolation method that estimates values at unobserved locations by weighting nearby observations using distance-based autocorrelation \u0026ndash; was implemented in four forms: simple, ordinary, universal, and empirical Bayesian. However, none of these studies incorporated covariates. Kriging studies were implemented mainly in ArcGIS (Esri, Redlands, CA, USA) software without the incorporation of covariates, and in most cases the resolution of these predictions was not reported. This analysis was also purely spatial with no temporal component, and the vast majority (83.3%) were conducted in Ethiopia only.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eBayesian geostatistical studies interpolate coverage spatially or spatiotemporally, at 1 km or 5 km resolution. These studies employed covariates (n=21), and they were implemented primarily using R-INLA package (n = 20) and Markov Chain Monte Carlo approaches (n=2). These studies typically assumed an independent, identically distributed non-spatial error term (n=9), a Matern covariance function (n=18), a Euclidean distance function (n=8), and/or defined the spatial decay (distance beyond which there is negligible spatial correlation) using an exponential function (n=5). Most also reported validation metrics, including information criteria and cross-validation statistics. A key advantage of Bayesian geostatistical methods is their ability to quantify the uncertainty associated with predictions. Most of these studies assessed uncertainty using 95% credible intervals, standard deviations, or non-exceedance probabilities. Most MBG studies were conducted in Nigeria (n=14) and only three in Ethiopia.\u003c/p\u003e\n\u003cp\u003eMachine learning\u003c/p\u003e\n\u003cp\u003eOnly seven studies (8%) used ML algorithms to estimate vaccination coverage. These were categorized as additive models (Generalized Additive Models), which combine linear and non-linear smoothing functions between predictors and the outcome, allowing flexible modelling of complex relationships; tree-based models (Random forests, Decision Tree, Gradient Boosting Model, Boosted regression trees), which recursively construct decision trees from input features, with each branch representing a decision that leads to a final prediction; regularization models (Least Absolute Shrinkage and Selection Operator -LASSO, ridge regression), which introduce constraints to prevent overfitting and improve generalizability of the model; neural networks (Multi-Layer Perceptron), which learn complex patterns through interconnected layers of nodes; and instance-based models (nearest neighbour), which rely on specific instances in the training data to make predictions (Table 2).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTools and reproducibility\u003c/p\u003e\n\u003cp\u003eTo model or map vaccination coverage, R (137) and STATA (138) were the two dominant statistical software used in 49 and 39 studies, respectively (Supplementary file 1 table S14). Desktop GIS software and specialized mapping tools were also employed with ArcGIS (Esri, Redlands, CA, USA) (n=38) and QGIS (n=6) (\u003cspan lang=\"EN-US\"\u003e(139)\u003c/span\u003e commonly used. These tools were primarily used for interpolation, data management and visualization of vaccination coverage data. Supplementary file 1 table S13 shows a summary of the tools and software used in the studies. Notably, to support reproducibility and transparency of the findings, only seven studies provided links to the code used to model vaccination coverage.\u003c/p\u003e\n\u003cp\u003eModelled outputs, limitations and recommendations\u003c/p\u003e\n\u003cp\u003eAggregation units, spatial and temporal resolution\u003c/p\u003e\n\u003cp\u003eOf the 86 modelling studies, 26 conducted spatial autocorrelation and clustering analysis without estimating coverage, hence they did not provide any spatial resolution of outputs. Across the remaining 60 studies that modelled vaccine coverage, spatial resolution of the outputs was presented as gridded surfaces only (high resolution) (n=7), at administrative units only (n=31) or as both gridded surfaces and at administrative units (n=22). Studies providing estimates as gridded surfaces used spatial resolutions of 1km (n=16), 5km (n=9), and 10km (one multi-country study), while two studies produced estimates at both 1km and 5km. One study did not report the spatial resolution. Overall, these studies predominantly employed geostatistical methods and covered 11 countries. The 22 studies that aggregated estimates to administrative units did so at varying levels of granularity: the first administrative level (region, provinces or states; n=4), at the second administrative unit (districts, counties or local government areas (LGAs) n=16) or at finer units such as cities in Niger and wards in Nigeria (Supplementary file 1 tables S15 \u0026amp; S16).\u003c/p\u003e\n\u003cp\u003eRegarding temporal resolution, of the 86 modelling studies, 27 produced annual estimates comprising of 18 country-level, 8 multi-country and 1 subnational level study. The remaining 59 studies reported estimates for a single point only (year).\u003c/p\u003e\n\u003cp\u003eReported limitations\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eData limitations\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eIssues related to vaccination data quality dominated the reported limitations. Since most studies relied on survey-based vaccination data, data quality limitations were mainly due to recall bias, non-response bias, and displacement of GPS coordinates. Studies using routine administrative data instead highlighted inaccuracies in the population denominators (the target population of children aged under five years who are due or eligible for a specific vaccine according to the national immunization schedule) used to estimate vaccination coverage. Data scarcity, incompleteness, or unavailability, particularly in hard-to-reach or conflict-affected areas, was another major challenge, directly affecting the representativeness of vaccination data. Additional constraints included the low spatial resolution of survey data due to aggregation at administrative levels (e.g., districts), the use of outdated vaccination data, cross-section nature of survey data and the inability to account for immunization delivered through SIAs.\u003c/p\u003e\n\u003cp\u003eChallenges with covariate data were mainly exclusion of key contextual covariates, such as supply-side factors like vaccine stocks, conflict exposure and migration in modelling, due to their unavailability or outdatedness. Additionally, use of covariates from multiple sources caused difficulties harmonizing multiple data sources.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eMethodological limitations\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eUncertainty in coverage estimates was not consistently reported, as some studies reported point estimates without associated confidence or credible intervals, limiting interpretation of the estimates. The ability to assess temporal patterns in coverage was also restricted because most analyses relied on single survey rounds or lacked longitudinal vaccination data.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eStudies that aggregated results to higher administrative units (e.g., provinces, districts) noted that this may have masked subnational heterogeneities. Some limitations were specific to the methods used, for instance, kriging performed poorly in unsampled areas, while SaTScan \u0026ndash; which relies on circular scanning windows \u0026ndash; may have missed irregularly shaped clusters. Studies using travel-time-based accessibility models cited oversimplification of real-world conditions by not accounting for weather conditions, transport availability, or traffic congestion. More details of data and methodological limitations are provided in Supplementary file 1 table S17.\u003c/p\u003e\n\u003cp\u003eReported recommendations\u003c/p\u003e\n\u003cp\u003eStudies emphasized improving vaccination data by using up-to-date datasets (including administrative boundaries, health facilities, vaccination status, and population distribution) and incorporating additional covariates such as vaccine hesitancy, facility readiness, and care-seeking behaviour. Standardized data collection guidelines, rigorous data quality checks, and triangulation of multiple sources \u0026ndash; including census data, satellite-derived population estimates, and age-stratified population data \u0026ndash; were recommended to enhance accuracy, account for seasonal migration, and reduce uncertainties in coverage estimates. Integrating data from multiple cross-sectional surveys and including qualitative information was also suggested to better help interpret coverage patterns, while strengthening the completeness and quality of routine immunization data was highlighted as critical for continuous monitoring, particularly at the subnational level. Methodological considerations to improve vaccination coverage modelling included subnational analyses to capture local heterogeneities, provision of model uncertainty to enhance interpretability and reliability of estimates, and continuous model refinement to include missing covariates and address geographic or urban-rural variability.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eUse of joint modelling frameworks to triangulate routine and survey vaccination data, as well as ML methods for automated feature extraction were also recommended. In addition, some studies proposed using online dashboards to facilitate interpretation and dissemination of results. Finally, continuous monitoring and evaluation of vaccination coverage, particularly in low-coverage areas, was recommended to track trends, inform interventions, and support adaptive modelling strategies over time. See Supplementary file 1 table S18 \u0026amp; S19 for more details. \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eDespite concerted efforts aimed at achieving full childhood immunization by key actors like GAVI and UNICEF (20,21,140), pockets of ZD children persist in LLMICs (141). Effectively reaching and vaccinating these children requires their accurate identification using spatially resolved data and robust geospatial methods. Our scoping review identified 102 studies conducted across 68 LLMICs that outlined relevant spatial data, spatial or spatial-statistical methods used to estimate childhood vaccine coverage, and the associated gaps. Nine in ten studies were conducted in just three countries (Ethiopia, Nigeria, or India), likely due to the high ZD burden in these countries (2). Additionally, 70% were published between 2021 and 2025, indicating that much of the evidence has emerged only recently.\u003c/p\u003e\n\u003cp\u003eRegardless of the growing recognition of the value of geospatial approaches for immunisation programming, our review shows a complex and uneven landscape. The mapping of ZD children, while gaining traction, remains limited in scope and uptake relative to its demonstrated utility for identifying pockets of under-immunisation. Definitions of ZD vary considerably across studies, undermining comparability and policy utility of findings. Data sources are heavily skewed toward household surveys, with routine administrative data remaining largely underused despite its potential for continuous, granular monitoring. Most studies rely on simpler exploratory methods without uncertainty-quantified estimates that operational microplanning demands, and spatiotemporal frameworks remain underused, limiting the capacity to track vaccination trends and link fluctuations to programmatic drivers. Below, we discuss these each point and its implications for the identification of ZD children in LLMICs.\u003c/p\u003e\n\u003cp\u003eOnly a fifth of the studies focused on mapping ZD prevalence indicating limited uptake of geospatial approaches for targeting ZD children, despite growing evidence of their ability to detect small pockets of under-immunization (50,55). This could be due to technical and structural barriers; ZD clusters tend to be rare, spatially heterogeneous, and often located in marginalized settings (26,50). Identifying them requires fine-resolution spatial data, advanced analyses, and cross-sector collaboration, which remain unevenly available in many LLMICs. Due to this complexity, such modelling studies are led by a handful of well-resourced academic groups, e.g., IHME local burden of disease and Worldpop\u0026rsquo;s VaxPop project \u003cspan class=\"CommentReference1\"\u003e(137,138)\u003c/span\u003e. Limited translation of ZD estimates in routine country planning, where administrative unit averages still dominate may further discourage adoption of geospatial approaches (49,144)\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAcross the studies, ZD was conceptualised as: i) not receiving any vaccine (66\u0026ndash;68), or ii) missing DTP1 (6,10,24,50,54\u0026ndash;65). Inconsistent definitions of ZD capture different populations which may lead to over or under-estimation of ZD prevalence and subsequently, important consequences on policy and programmatic actions such as inefficient strategies for reaching the most disadvantaged children (130). For example, globally, 14.2% of children are DTP1-ZD and 7.5% are completely ZD (no single vaccine) (145). Despite recommended use of DTP1 for operational purposes (4), our review emphasizes the need to further standardize and harmonise definitions of ZD, and appropriate target age-groups. Inconsistent definitions and target age-groups lead to incomparability of evidence limiting its usefulness for global policy formulation, prioritization and benchmarking progress. In addition, inconsistent definitions can yield divergent interpretations of drivers of ZD and weaken assessments of equity gradients across settings. On the other hand, definitional variation is not without rationale. DTP1 definition carries clear operational justification: administrative data capture DTP1 non-receipt continuously and at facility level, enabling real-time monitoring at a granularity that survey-based definitions cannot match, and at population level DTP1 coverage tracks closely with rates of complete non-vaccination (146,147).\u0026nbsp;Conversely, the broader definition of receiving no vaccine may better capture the most marginalised children in settings where even partial immunisation is rare (146,147).\u003c/p\u003e\n\u003cp\u003eThe main source of both vaccination and covariate data was household survey data, particularly DHS (\u003cstrong\u003eSupplementary file 1 table S8\u003c/strong\u003e). These surveys are nationally representative, standardized, less vulnerable to reporting incompleteness or treatment-seeking biases with well-defined population denominators and detailed socio-demographic indicators. However, they do not capture subnational heterogeneities due to low spatial resolution (missing most vulnerable populations), are cross-sectional, and are conducted less frequently \u0026ndash; every three to five years (148). The recent defunding of the DHS Program further threatens the tracking of health indicators in LLMICs including ZD (35). In contrast, there was very low use of routine administrative vaccination data despite its increased availability in the last decade through DHIS2 (149). Routine data is continuous, timely and geographically disaggregated and can enable continuous ZD monitoring and localized decision-making. However, it is prone to incompleteness, outliers, and inaccurate population denominators for coverage computation (62,112,150). Further, as routine data are collected at the health facility level, estimating denominators and health facility catchment areas is challenging (151), particularly in settings with high population mobility, inaccurate population estimates and low healthcare utilization rates. Therefore, making use of routine data requires advanced methods, which may explain why studies preferred survey-based data (114). This highlights a need to address systemic and structural gaps to improve collection and reporting of HMIS data (152), use of rapid and targeted assessment tools such as lot quality assurance sampling and micro-planning to pinpoint low coverage areas (153), and triangulation of survey and routine data to leverage their differing strengths (151).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOnly five studies used both survey and routine data, treating them separately (e.g., using DHS coverage to adjust DHIS2 denominators) rather than jointly modelling them. There is need for joint integration of routine and survey data in a modelling framework to maximise the complementary strengths of both data types. Such integrated frameworks have proven superior in other applications, for instance in mapping malaria incidence (151), by improving estimation accuracy and revealing hidden inequities that single-source models miss. Establishing integrated routine\u0026ndash;survey modelling pipelines could represent a high-impact priority for the next generation of ZD mapping, enabling both frequent monitoring and robust estimation in data-sparse areas.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe identified six categories of the covariates used, with demographic, health-system, and physical-environment factors being the most applied, consistent with previous findings that highlight demographic and healthcare access factors as key barriers to vaccination among ZD communities (154). Still, there is a significant gap in the covariates used to capture fragile, marginalized, or hard-to-reach populations \u0026ndash; such as conflict-affected, remote, or highly mobile populations (e.g., nomadic pastoralists, refugees, internally displaced persons), and urban slum dwellers. This was mainly due to data unavailability or scarcity in these settings (26,54,55). Given that ZD children are disproportionately concentrated in marginalized and fragile settings, the lack of covariates capturing these contexts limits the ability to identify populations at greatest risk of ZD. Studies modelling ZD prevalence should therefore incorporate at least one covariate reflecting marginalization, depending on the study context and scope. Collaborative efforts by data agencies, local health authorities and international partners are needed to expand covariate data availability to avoid systematically missing the most disadvantaged populations.\u003c/p\u003e\n\u003cp\u003eWhile there were diverse modelling methods, most studies relied on simpler, exploratory approaches to examine autocorrelation and clustering of vaccination data. These methods do not interpolate estimates for unsampled locations and thus cannot generate gridded estimates. Gridded or continuous estimates are preferable, as they can be directly overlaid with high resolution population density maps for absolute estimates of ZD or other gridded estimate to identify overlapping dual or triple burdens (6,26). These exploratory approaches were implemented in mapping software (especially ArcGIS (Esri, Redlands, CA, USA)) which have a graphical user interface and require less technical expertise and computational resources than advanced modelling approaches. However, these descriptive clustering approaches alone are insufficient for operational microplanning, which increasingly requires small area estimates with quantified uncertainty for prioritization and monitoring.\u003c/p\u003e\n\u003cp\u003eMore complex modelling approaches including SAE, MBG and ML are better suited to ZD estimation as they incorporate covariates, generate uncertainty and produce outputs at operationally meaningful scales (155,156). The resolution at which the modelled outputs are produced matters. Gridded estimates are particularly valuable for hotspot identification and the targeting of interventions, as they can be directly overlaid with high-resolution population data to identify absolute numbers of ZD children and areas of overlapping deprivation (67,125). These gridded outputs can then be aggregated to subnational administrative units that align with the scales at which immunisation programmes are planned and evaluated (114,125), support decentralised resource targeting, and reveal subnational inequities masked by national averages (49). Aggregation and produces outputs more interpretable for health managers, while enabling alignment with global policy benchmarks such as GVAP, IA2030) (122).\u003c/p\u003e\n\u003cp\u003eYet, uptake of these methods remains low in low-resource settings due to steep requirements for expertise and computing resources. National programs in LLMICs often lack sufficient local capacity and often outsource complex analyses, reducing ownership and timeliness evidence for decision-making (157). Compounding this, very few studies provided analytical code, limiting reproducibility and the ability of country teams to adapt and reuse methods. Alternative user-friendly web tools such as Maplaria (158), MBGapp (159), and sae4health (160) address these barriers by automating pipelines, hosting computations remotely, and requiring no coding or advanced statistical knowledge. Bridging this gap through investment in local capacity, user-friendly analytical tools, and stronger partnerships between academic groups and national programmes is essential to strengthen country ownership of geospatial evidence and improve the timeliness of targeting decisions for zero-dose programming.\u003c/p\u003e\n\u003cp\u003eFew studies used AI approaches, indicative of a slow uptake of these methods in this domain. While a comparative study showed that geostatistical models slightly outperform ML models in estimating vaccine coverage (29),\u0026nbsp;it is still necessary to explore the full potential of AI approaches, especially as the availability of geospatial big data and satellite-derived covariates increases. These models have the capacity to handle large, multidimensional datasets, such as DHS and routine HMIS data, enabling integration of variables (e.g., demographic, health system, environmental) and recognition of non-linear patterns in the variables (47). They can also improve the efficiency of analytical tasks, for instance, using Large Language Models (LLMs) to automate feature extraction (161,162).\u003c/p\u003e\n\u003cp\u003eOnly 31% of the studies incorporated a spatiotemporal framework, reflecting the complexity of spatiotemporal modelling and the scarcity of longitudinal data, as most analyses relied on cross-sectional survey data. Temporal estimates are critical for understanding vaccination trends over time and linking fluctuations to events or seasonal factors. Where longitudinal routine data are available, spatiotemporal approaches bridge microplanning needs and dynamic monitoring of immunization performance. Promoting the availability and use of longitudinal routine data, alongside advancing spatiotemporal modelling approaches is essential to enhance temporal analysis and prediction of ZD prevalence. The role of infection prevalence, transmission dynamics, and disease burden remains insufficiently integrated into studies assessing correlations with ZD children, potentially biasing interpretation of observed associations.\u003c/p\u003e\n\u003cp\u003eTo advance this field, our review characterises the landscape of spatial vaccine coverage modelling in LLMICs, highlighting persistent data and methodological limitations that constrain the accuracy, equity, and comparability of estimates. Hard-to-reach populations are systematically excluded from analysis due to the lack of both reliable vaccination data and covariates capturing key contextual factors such as conflict exposure, poverty, remoteness and migration. This leads to model misspecification in the settings where ZD burden is highest. Heavy reliance on household surveys, with limited triangulation with DHIS2 data, further restricts the effective use of routine data, which are themselves undermined by incomplete reporting and inaccurate population denominators.\u003c/p\u003e\n\u003cp\u003eAddressing these challenges will require harmonised definitions of ZD children, improved integration of survey and routine data through joint modelling frameworks, and the incorporation of non-traditional data sources such as satellite imagery, mobility data, and conflict databases. It will also require strengthening of data collection in underserved settings through approaches such as community health worker enumeration and rapid assessments (\u003cstrong\u003eSupplementary File 3)\u003c/strong\u003e.\u003c/p\u003e\n\u003ch2\u003eStrengths and Limitations\u003c/h2\u003e\n\u003cp\u003eThis review synthesizes evidence from a broad range of geospatial studies focused on assessing childhood vaccination coverage, providing a consolidated understanding the gaps in data and sources, methods, and structural limitations of studies already available in the literature. To our knowledge, it is the first review to examine ZD prevalence through a geospatial lens, offering a unique contribution to the field and helping clarify a potential roadmap on how spatial data and methods can be applied to identify ZD children. However, the review only included articles published in English, which may have resulted in the exclusion of relevant studies published in other languages. In addition, the included studies varied in ZD definitions, denominators, spatial scales, and modelling approaches, which limited direct comparability of estimates across settings.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eOur review emphasizes the critical role that geospatial data and methods play in modelling and mapping ZD prevalence across diverse settings. It also highlights substantial gaps and limitations in the data and methods used, offering insights into where improvements are needed to better identify ZD children. Future research should prioritize triangulating routine immunization and household survey data while placing greater emphasis on understudied populations such as nomadic groups, refugees, urban slum residents, and conflict-prone areas. In addition, studies should move beyond describing spatial patterns through clustering and similar methods and progress toward generating monitoring tools and predictive estimates of ZD prevalence at detailed spatiotemporal scales, which are more actionable for targeted interventions and policymaking. Overall, this review provides a foundational reference for future ZD surveillance and modelling work as well as for policymakers seeking to integrate geospatial evidence into programming and monitoring. The covariates and methodological approaches identified across studies offer a practical starting point for researchers, while the documented gaps and recommendations can help guide efforts to strengthen, harmonize, and streamline geospatial modelling of ZD populations. Ultimately, improving standardization, strengthening routine data and analytic capacity, and expanding research into fragile and underserved contexts will be essential to ensure that geospatial evidence translates into equitable immunization gains.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eThe REACHOUT Consortium\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCarlo Federici, Samuel Muhula, Jeanine Condo, Fabrizio Tediosi, Bolanle Oyeledun, Piero Poletti, Francesco Menegale, Manuela De Allegri, John Kutna, Yvonne Opanga, Herbert Barasa, Joan Mboga, Anne Gitimu, Caroline Mudereri, Gashaija Absolomon, Felix Rubuga Kitema, Hassan Sibomana, Olivier Wane, Jean Damascene Hagenimana, Grace Kabanyana, Jean de Dieu Hakizimana, Emma Clarke-Deelder, Amit Aryal, Obioma Ezebuka, Francis Ogirima, Oluwatosin Oladokun, Abimbola Phillips, David Udanwojo, Michael Omobhude, Carlos Felipe Balmaceda, Alessia Melegaro, Maria Cucinello, Aleksandra Torbica, Vittoria Offeddu, Ankit Shanker, Phidelis Wamalwa, Kavita Singh, Swati Srivastava\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCRediT authorship contribution statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization: PMM, AMN, CF\u003c/p\u003e\n\u003cp\u003eMethodology: PMM, AMN\u003c/p\u003e\n\u003cp\u003eSoftware: PMM, AMN, MMM\u003c/p\u003e\n\u003cp\u003eValidation: PMM, AMN\u003c/p\u003e\n\u003cp\u003eFormal analysis: PMM, AMN, MMM\u003c/p\u003e\n\u003cp\u003eInvestigation:\u0026nbsp;PMM, AMN\u003c/p\u003e\n\u003cp\u003eResources: PMM\u003c/p\u003e\n\u003cp\u003eData Curation: PMM, AMN, MMM\u003c/p\u003e\n\u003cp\u003eWriting - Original Draft: PMM, AMN\u003c/p\u003e\n\u003cp\u003eWriting - Review \u0026amp; Editing, PMM, AMN, MMM, SS, CM, FRK, YO, EC, FM, AM, AG, JC, FK, LB, JIB, PP, AT, CF, AM\u003c/p\u003e\n\u003cp\u003eVisualization: PMM, AMN\u003c/p\u003e\n\u003cp\u003eSupervision: PMM\u003c/p\u003e\n\u003cp\u003eProject administration: PMM, CF\u003c/p\u003e\n\u003cp\u003eFunding acquisition:\u0026nbsp;PMM, SS, JC, PP, AT, AM, CF\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNot applicable.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of Generative AI and AI-assisted technologies in the writing process\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eChatGPT- text editing\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe work was funded by the European Union under The Global Health EDCTP3 Joint Undertaking (GH EDCTP3 JU) \u0026ndash; Grant Agreement number 101159477. Views and opinions expressed are however, those of the authors only and do not necessarily reflect those of the European Union or GH EDCTP3. Neither the European Union nor the granting authority can be held responsible for them\u003cstrong\u003e.\u0026nbsp;\u003c/strong\u003ePMM is supported by Fonds voor Wetenschappelijk Onderzoek (FWO) for his Senior Postdoctoral Fellowship (Grant number: 1201925N)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of competing interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank Rehema Ouko (Department of Public Health, Institute of Tropical Medicine (ITM), Antwerp, Belgium) and Dr. Aliki Christou (Department of Public Health, Institute of Tropical Medicine (ITM), Antwerp, Belgium) for their assistance in obtaining access to relevant articles used in this review.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study is based on secondary analysis of previously published studies.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eWHO. Immunization [Internet]. [cited 2025 Nov 25]. Available from: https://www.afro.who.int/health-topics/immunization\u003c/li\u003e\n\u003cli\u003eWUENIC Trends [Internet]. [cited 2026 Jan 27]. Available from: https://worldhealthorg.shinyapps.io/wuenic-trends/\u003c/li\u003e\n\u003cli\u003eGAVI. The Zero-Dose Child: Explained [Internet]. [cited 2025 Nov 25]. Available from: https://www.gavi.org/vaccineswork/zero-dose-child-explained\u003c/li\u003e\n\u003cli\u003eGAVI. Zero-dose children and missed communities [Internet]. [cited 2025 Nov 25]. Available from: https://www.gavi.org/our-alliance/strategy/phase-5-2021-2025/equity-goal/zero-dose-children-missed-communities\u003c/li\u003e\n\u003cli\u003eCorr\u0026ecirc;a GC, Uddin MJ, Wahed T, Oliveras E, Morgan C, Kamya MR, et al. Measuring Zero-Dose Children: Reflections on Age Cohort Flexibilities for Targeted Immunization Surveys at the Local Level. Vaccines (Basel). 2024 Feb 1;12(2). doi:10.3390/vaccines12020195\u003c/li\u003e\n\u003cli\u003eUtazi CE, Chan HMT, Olowe I, Wigley A, Tejedor-Garavito N, Cunningham A, et al. A zero-dose vulnerability index for equity assessment and spatial prioritization in low- and middle-income countries. Spatial Statistics. Elsevier B.V.; 2023. doi:10.1016/j.spasta.2023.100772\u003c/li\u003e\n\u003cli\u003eArambepola R, Yang Y, Hutchinson K, Mwansa FD, Doherty JA, Bwalya F, et al. Using geospatial models to map zero-dose children: Factors associated with zero-dose vaccination status before and after a mass measles and rubella vaccination campaign in Southern province, Zambia. BMJ Glob Health. 2021 Dec 30;6(12). doi:10.1136/bmjgh-2021-007479\u003c/li\u003e\n\u003cli\u003eRachlin A, Danovaro-Holliday MC, Murphy P, Sodha S V., Wallace AS. Routine Vaccination Coverage \u0026mdash; Worldwide, 2021. MMWR Morb Mortal Wkly Rep. 2022;71(44). doi:10.15585/mmwr.mm7144a2\u003c/li\u003e\n\u003cli\u003eUtazi CE, Thorley J, Alegana VA, Ferrari MJ, Takahashi S, Metcalf CJE, et al. Mapping vaccination coverage to explore the effects of delivery mechanisms and inform vaccination strategies. Nat Commun. 2019 Dec 1;10(1). doi:10.1038/s41467-019-09611-1 PubMed PMID: 30967543.\u003c/li\u003e\n\u003cli\u003eMbunga BK, Liu PY, Bangelesa F, Mafuta E, Dalau NM, Egbende L, et al. Zero-Dose Childhood Vaccination Status in Rural Democratic Republic of Congo: Quantifying the Relative Impact of Geographic Accessibility and Attitudes toward Vaccination. Vaccines (Basel). 2024 Jun 1;12(6). doi:10.3390/vaccines12060617\u003c/li\u003e\n\u003cli\u003eArambepola R, Yang Y, Hutchinson K, Mwansa FD, Doherty JA, Bwalya F, et al. Using geospatial models to map zero-dose children: Factors associated with zero-dose vaccination status before and after a mass measles and rubella vaccination campaign in Southern province, Zambia. BMJ Glob Health. 2021 Dec 30;6(12). doi:10.1136/bmjgh-2021-007479\u003c/li\u003e\n\u003cli\u003eFMoHSW FM of H and SW of N, NPC/Nigeria NPC, ICF. Nigeria Demographic and Health Survey 2024 [Internet]. 2025 Oct [cited 2025 Nov 25]. Available from: https://dhsprogram.com/publications/publication-FR395-DHS-Final-Reports.cfm\u003c/li\u003e\n\u003cli\u003eGAVI. Turning the corner: IA2030 Global Report 2023. 2023.\u003c/li\u003e\n\u003cli\u003eMurhekar M V., Kumar MS. Reaching zero-dose children in India: progress and challenges ahead. The Lancet Global Health. Elsevier Ltd; 2021. p. e1630\u0026ndash;1. doi:10.1016/S2214-109X(21)00406-X PubMed PMID: 34798011.\u003c/li\u003e\n\u003cli\u003eMalande OO, Meyer JC, Adamu VE, Munube D, Muhindo R, Kigen B, et al. Zero-dose or under immunized: Where should the emphasis lie? Journal of the African Society for Paediatric Infectious Diseases. 2024 Dec 30;3:1\u0026ndash;9. doi:10.15641/jafspidvol3pp1-9/1732\u003c/li\u003e\n\u003cli\u003eProgress and challenges with Achieving Universal Immunization Coverage [Internet]. [cited 2025 Nov 21]. Available from: https://www.who.int/publications/m/item/progress-and-challenges\u003c/li\u003e\n\u003cli\u003eRainey JJ, Watkins M, Ryman TK, Sandhu P, Bo A, Banerjee K. Reasons related to non-vaccination and under-vaccination of children in low and middle income countries: Findings from a systematic review of the published literature, 1999-2009. Vaccine. 2011. p. 8215\u0026ndash;21. doi:10.1016/j.vaccine.2011.08.096 PubMed PMID: 21893149.\u003c/li\u003e\n\u003cli\u003eJani J V., De Schacht C, Jani I V., Bjune G. Risk factors for incomplete vaccination and missed opportunity for immunization in rural Mozambique. BMC Public Health. 2008;8. doi:10.1186/1471-2458-8-161 PubMed PMID: 18485194.\u003c/li\u003e\n\u003cli\u003eGAVI. IMMUNIZATION AGENDA 2030: A Global Strategy to Leave No One Behind [Internet]. 2020 [cited 2025 Aug 22]. Available from: https://www.who.int/teams/immunization-vaccines-and-biologicals/strategies/ia2030\u003c/li\u003e\n\u003cli\u003eGAVI. Phase 5 (2021\u0026ndash;2025) [Internet]. [cited 2025 Nov 20]. Available from: https://www.gavi.org/our-alliance/strategy/phase-5-2021-2025\u003c/li\u003e\n\u003cli\u003ePhase 6 (2026\u0026ndash;2030) [Internet]. [cited 2025 Nov 24]. Available from: https://www.gavi.org/our-alliance/strategy/phase-6-2026-2030#vision\u003c/li\u003e\n\u003cli\u003eWHO. IA2030 IG 2.1: Number of zero dose children [Internet]. [cited 2025 Nov 25]. Available from: https://www.who.int/data/gho/indicator-metadata-registry/imr-details/7792\u003c/li\u003e\n\u003cli\u003eUnited Nations. The Sustainable Development Goals Report [Internet]. 2025 [cited 2025 Aug 12]. Available from: https://unstats.un.org/sdgs/report/2025/The-Sustainable-Development-Goals-Report-2025.pdf\u003c/li\u003e\n\u003cli\u003eUtazi CE, Aheto JMK, Wigley A, Tejedor-Garavito N, Bonnie A, Nnanatu CC, et al. Mapping the distribution of zero-dose children to assess the performance of vaccine delivery strategies and their relationships with measles incidence in Nigeria. Vaccine. 2023 Jan 4;41(1):170\u0026ndash;81. doi:10.1016/j.vaccine.2022.11.026 PubMed PMID: 36414476.\u003c/li\u003e\n\u003cli\u003eUtazi CE, Thorley J, Alegana VA, Ferrari MJ, Takahashi S, Metcalf CJE, et al. High resolution age-structured mapping of childhood vaccination coverage in low and middle income countries. Vaccine. 2018 Mar 14;36(12):1583\u0026ndash;91. doi:10.1016/j.vaccine.2018.02.020 PubMed PMID: 29454519.\u003c/li\u003e\n\u003cli\u003eWigley A, Lorin J, Hogan D, Utazi CE, Hagedorn B, Dansereau E, et al. Estimates of the number and distribution of zero-dose and under-immunised children across remote-rural, urban, and conflict-affected settings in low and middle-income countries. PLOS Global Public Health. 2022 Oct 1;2(10 October). doi:10.1371/journal.pgph.0001126\u003c/li\u003e\n\u003cli\u003eGAVI. IRMMA framework \u0026ndash; guidance on use of Gavi support to reach zero-dose children and missed communities [Internet]. 2021 [cited 2025 Aug 26]. Available from: https://immunizationeconomics.org/recent-activity/2021/8/1/irmma-framework-guidance-on-use-of-gavi-support-to-reach-zero-dose-children-and-missed-communities/\u003c/li\u003e\n\u003cli\u003eUtazi CE, Thorley J, Alegana VA, Ferrari MJ, Takahashi S, Metcalf CJE, et al. High resolution age-structured mapping of childhood vaccination coverage in low and middle income countries. Vaccine. 2018 Mar 14;36(12):1583\u0026ndash;91. doi:10.1016/j.vaccine.2018.02.020 PubMed PMID: 29454519.\u003c/li\u003e\n\u003cli\u003eUtazi CE, Yankey O, Chaudhuri S, Olowe ID, Danovaro-Holliday MC, Lazar AN, et al. Geostatistical and machine learning approaches for high-resolution mapping of vaccination coverage. Spat Spatiotemporal Epidemiol. 2025 Aug 1;54:100744. doi:10.1016/J.SSTE.2025.100744 PubMed PMID: 40935503.\u003c/li\u003e\n\u003cli\u003eTatem AJ. Innovation to impact in spatial epidemiology Sally Blower; Gerardo Chowell. BMC Med. 2018 Nov 14;16(1). doi:10.1186/s12916-018-1205-5 PubMed PMID: 30424754.\u003c/li\u003e\n\u003cli\u003eKirby RS, Delmelle E, Eberth JM. Advances in spatial epidemiology and geographic information systems. Ann Epidemiol. 2017;27(1). doi:10.1016/j.annepidem.2016.12.001\u003c/li\u003e\n\u003cli\u003eDelmelle EM, Desjardins MR, Jung P, Owusu C, Lan Y, Hohl A, et al. Uncertainty in geospatial health: challenges and opportunities ahead. Annals of Epidemiology. 2022. doi:10.1016/j.annepidem.2021.10.002\u003c/li\u003e\n\u003cli\u003eMorrison CN, Mair CF, Bates L, Duncan DT, Branas CC, Bushover BR, et al. Defining Spatial Epidemiology: A Systematic Review and Re-orientation. Epidemiology. 2024. doi:10.1097/EDE.0000000000001738\u003c/li\u003e\n\u003cli\u003eCorsi DJ, Neuman M, Finlay JE, Subramanian S V. Demographic and health surveys: A profile. Int J Epidemiol. 2012 Dec;41(6):1602\u0026ndash;13. doi:10.1093/ije/dys184 PubMed PMID: 23148108.\u003c/li\u003e\n\u003cli\u003eKhaki JJ, Molenaar J, Karki S, Olal E, Straneo M, Mosuse MA, et al. When health data go dark: the importance of the DHS Program and imagining its future. BMC Med. 2025 Dec 1;23(1). doi:10.1186/s12916-025-04062-6 PubMed PMID: 40275318.\u003c/li\u003e\n\u003cli\u003eMacharia PM, Beňov\u0026aacute; L, Ray N, Semaan A, Musau MM, Kipterer JK, et al. Putting health facilities on the map: a renewed call to create geolocated, comprehensive, updated, openly licensed dataset of health facilities in sub-Saharan African countries. BMC Medicine 2025 23:1. 2025 Apr 7;23(1):211-. doi:10.1186/S12916-025-04023-Z PubMed PMID: 40197252.\u003c/li\u003e\n\u003cli\u003ePopulation Reference Bureau. The DHS Program - Evolution of The DHS Program [Internet]. [cited 2025 Nov 25]. Available from: https://dhsprogram.com/data/visualizations/Evolution-of-The-DHS-Program.cfm\u003c/li\u003e\n\u003cli\u003eDHIS2. DHIS2 In Action [Internet]. [cited 2025 Nov 25]. Available from: https://dhis2.org/in-action/#map\u003c/li\u003e\n\u003cli\u003eWorld Bank Country and Lending Groups \u0026ndash; World Bank Data Help Desk [Internet]. [cited 2026 Feb 3]. Available from: https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups\u003c/li\u003e\n\u003cli\u003eFavin M, Steinglass R, Fields R, Banerjee K, Sawhney M. Why children are not vaccinated: A review of the grey literature. International Health. 2012. p. 229\u0026ndash;38. doi:10.1016/j.inhe.2012.07.004 PubMed PMID: 24029668.\u003c/li\u003e\n\u003cli\u003eTauil M de C, Sato APS, Waldman EA. Factors associated with incomplete or delayed vaccination across countries: A systematic review. Vaccine. Elsevier Ltd; 2016. p. 2635\u0026ndash;43. doi:10.1016/j.vaccine.2016.04.016 PubMed PMID: 27109562.\u003c/li\u003e\n\u003cli\u003eGaladima AN, Zulkefli NAM, Said SM, Ahmad N. Factors influencing childhood immunisation uptake in Africa: a systematic review. BMC Public Health. 2021 Dec 1;21(1). doi:10.1186/s12889-021-11466-5 PubMed PMID: 34320942.\u003c/li\u003e\n\u003cli\u003eBangura JB, Xiao S, Qiu D, Ouyang F, Chen L. Barriers to childhood immunization in sub-Saharan Africa: A systematic review. BMC Public Health. BioMed Central; 2020. doi:10.1186/s12889-020-09169-4 PubMed PMID: 32664849.\u003c/li\u003e\n\u003cli\u003eGuzman-Holst A, DeAntonio R, Prado-Cohrs D, Juliao P. Barriers to vaccination in Latin America: A systematic literature review. Vaccine. Elsevier Ltd; 2020. p. 470\u0026ndash;81. doi:10.1016/j.vaccine.2019.10.088 PubMed PMID: 31767469.\u003c/li\u003e\n\u003cli\u003eBelt RV, Abdullah S, Mounier-Jack S, Sodha S V., Danielson N, Dadari I, et al. Improving Equity in Urban Immunization in Low- and Middle-Income Countries: A Qualitative Document Review. Vaccines. Multidisciplinary Digital Publishing Institute (MDPI); 2023. doi:10.3390/vaccines11071200\u003c/li\u003e\n\u003cli\u003eLyons C, Nambiar D, Johns NE, Allorant A, Bergen N, Hosseinpoor AR. Inequality in Childhood Immunization Coverage: A Scoping Review of Data Sources, Analyses, and Reporting Methods. Vaccines. Multidisciplinary Digital Publishing Institute (MDPI); 2024. doi:10.3390/vaccines12080850\u003c/li\u003e\n\u003cli\u003eCrocker-Buque T, Mindra G, Duncan R, Mounier-Jack S. Immunization, urbanization and slums - A systematic review of factors and interventions. BMC Public Health. 2017 Jun 8;17(1). doi:10.1186/s12889-017-4473-7 PubMed PMID: 28595624.\u003c/li\u003e\n\u003cli\u003eMusuka G, Umar A umra, Dadari I, Moyo E, Mano O, Iradukunda PG, et al. Bridging immunization gaps in Sub-Saharan Africa: a narrative review of microplanning, geospatial, and machine learning approaches to reach zero-dose children and under-immunised children. Vaccine. 2026 Apr;79:128413. doi:10.1016/j.vaccine.2026.128413 PubMed PMID: 41806761.\u003c/li\u003e\n\u003cli\u003eReaching the unreached with life-saving vaccines through data science and geospatial technologies - World | ReliefWeb [Internet]. [cited 2026 Mar 18]. Available from: https://reliefweb.int/report/world/reaching-unreached-life-saving-vaccines-through-data-science-and-geospatial-technologies\u003c/li\u003e\n\u003cli\u003eBaptiste AEJ, Wagai J, Hahn\u0026eacute; S, Adeniran A, Koko RI, de Vos S, et al. High-Resolution Geospatial Mapping of Zero-Dose and Underimmunized Children Following Nigeria\u0026rsquo;s 2021 Multiple Indicator Cluster Survey/National Immunization Coverage Survey. Journal of Infectious Diseases. 2024 Jul 15;230(1):e131\u0026ndash;8. doi:10.1093/infdis/jiad476 PubMed PMID: 39052714.\u003c/li\u003e\n\u003cli\u003eTricco AC, Lillie E, Zarin W, O\u0026rsquo;Brien KK, Colquhoun H, Levac D, et al. PRISMA extension for scoping reviews (PRISMA-ScR): Checklist and explanation. Ann Intern Med. 2018 Oct 2;169(7):467\u0026ndash;73. doi:10.7326/M18-0850 PubMed PMID: 30178033.\u003c/li\u003e\n\u003cli\u003eNjogu A, Mutinda M, Macharia P. Geospatial data and methods for mapping zero-dose children in low- and middle-income countries: A scoping literature review protocol [Internet]. 2025 May 28. doi:10.17605/OSF.IO/2Z3XV\u003c/li\u003e\n\u003cli\u003ePeters MD, Godfrey CM, Mclnerney P, Soares CB, Khalil H, Parker D. The Joanna Briggs Institute Reviewers\u0026rsquo; Manual 2015. Methodology for JBI Scoping Reviews. 2015.\u003c/li\u003e\n\u003cli\u003eAlegana VA, Ticha JM, Mwenda JM, Katsande R, Gacic-Dobo M, Danovaro-Holliday MC, et al. Modelling the spatial variability and uncertainty for under-vaccination and zero-dose children in fragile settings. Sci Rep. 2024 Dec 1;14(1). doi:10.1038/s41598-024-74982-5 PubMed PMID: 39420047.\u003c/li\u003e\n\u003cli\u003eMosser JF, Gagne-Maynard W, Rao PC, Osgood-Zimmerman A, Fullman N, Graetz N, et al. Mapping diphtheria-pertussis-tetanus vaccine coverage in Africa, 2000\u0026ndash;2016: a spatial and temporal modelling study. The Lancet. 2019 May 4;393(10183):1843\u0026ndash;55. doi:10.1016/S0140-6736(19)30226-0 PubMed PMID: 30961907.\u003c/li\u003e\n\u003cli\u003eBiswas A, Tucker J, Bauhoff S. Performance of predictive algorithms in estimating the risk of being a zero-dose child in India, Mali and Nigeria. BMJ Glob Health. 2023 Oct 11;8(10). doi:10.1136/bmjgh-2023-012836\u003c/li\u003e\n\u003cli\u003eEdson Utazi C, Pannell O, Aheto JMK, Wigley A, Tejedor-Garavito N, Wunderlich J, et al. Assessing the characteristics of un- and under-vaccinated children in low- and middle-income countries: A multi-level cross-sectional study. PLOS Global Public Health. 2022 Apr 1;2(4 April). doi:10.1371/journal.pgph.0000244\u003c/li\u003e\n\u003cli\u003eAheto JMK, Olowe ID, Chan HMT, Ekeh A, Dieng B, Fafunmi B, et al. Geospatial Analyses of Recent Household Surveys to Assess Changes in the Distribution of Zero-Dose Children and Their Associated Factors before and during the COVID-19 Pandemic in Nigeria. Vaccines (Basel). 2023 Dec 1;11(12). doi:10.3390/vaccines11121830\u003c/li\u003e\n\u003cli\u003eShiferie F, Gebremedhin S, Andargie G, DelPizzo F, Belete K, Fenta TG. Spatial distribution of zero-dose children in Ethiopia: evidence for a targeted intervention from a large-scale cross-sectional evaluation survey. Front Pediatr. 2024;12. doi:10.3389/fped.2024.1337922\u003c/li\u003e\n\u003cli\u003eSaidu Y, Agbor VN, Di Mattei P, Nchinjoh SC, Edwidge NN, Njoh AA, et al. Development of a composite scoring system to rank communities at high risk of zero-dose children in Cameroon: A geospatial analysis. J Glob Health. 2023;13. doi:10.7189/jogh.13.04136 PubMed PMID: 37971948.\u003c/li\u003e\n\u003cli\u003eHaeuser E, Nguyen JQ, Rolfe S, Nesbit O, Fullman N, Mosser JF. Assessing Geographic Overlap between Zero-Dose Diphtheria\u0026ndash;Tetanus\u0026ndash;Pertussis Vaccination Prevalence and Other Health Indicators. Vaccines (Basel). 2023 Apr 1;11(4). doi:10.3390/vaccines11040802\u003c/li\u003e\n\u003cli\u003eMehmood M, Setayesh H, Siddiqi DA, Siddique M, Iftikhar S, Soundardjee R, et al. Prevalence, geographical distribution and factors associated with pentavalent vaccine zero dose status among children in Sindh, Pakistan: analysis of data from the 2017 and 2018 birth cohorts enrolled in the provincial electronic immunisation registry. BMJ Open. 2022 May 18;12(5). doi:10.1136/bmjopen-2021-058985 PubMed PMID: 35584879.\u003c/li\u003e\n\u003cli\u003eMutwadi AM, Madinga JN, Vanlerberghe V, Mbala PK, van der Sande MAB. Geospatial and multilevel clustering of zerodose children in Kikwit, Democratic Republic of the Congo in 2022. PLOS Global Public Health. 2024 Feb 29;4(2). doi:10.1371/journal.pgph.0002617\u003c/li\u003e\n\u003cli\u003eFullman N, Correa GC, Ikilezi G, Phillips DE, Reynolds HW. Assessing Potential Exemplars in Reducing Zero-Dose Children: A Novel Approach for Identifying Positive Outliers in Decreasing National Levels and Geographic Inequalities in Unvaccinated Children. Vaccines (Basel). 2023 Mar 1;11(3). doi:10.3390/vaccines11030647\u003c/li\u003e\n\u003cli\u003eJohri M, Rajpal S, Kim R, Subramanian S V. Small-area variation in child under-vaccination in India: a multilevel analysis of cross-sectional data from 36 states and Union Territories, 707 districts, and 22,349 small-area clusters. The Lancet Regional Health - Southeast Asia. 2025 Jan 1;32. doi:10.1016/j.lansea.2024.100504\u003c/li\u003e\n\u003cli\u003eTamir TT, Zegeye AF, Mekonen EG, Tekeba B, Ali MS, Gonete AT, et al. Prevalence, spatial variation and determinants of zero-dose children in Ethiopia: Spatial and multilevel analyses. Public Health. 2024 Nov 1;236:365\u0026ndash;72. doi:10.1016/j.puhe.2024.09.011 PubMed PMID: 39303624.\u003c/li\u003e\n\u003cli\u003eUtazi CE, Aheto JMK, Chan HMT, Tatem AJ, Sahu SK. Conditional probability and ratio-based approaches for mapping the coverage of multi-dose vaccines. Stat Med. 2022 Dec 20;41(29):5662\u0026ndash;78. doi:10.1002/sim.9586 PubMed PMID: 36129171.\u003c/li\u003e\n\u003cli\u003eAgimas MC, Asmamaw M, Hailu MK, Kidie T, Abuhay HW, Yismaw GA, et al. Geospatial mapping to assess the distribution and determinants of zero dose vaccination status hot spots among children in Ethiopia using EDHS 2019: Spatial and geographical weighted regression. PLoS One. 2024 Dec 1;19(12). doi:10.1371/journal.pone.0312610 PubMed PMID: 39739825.\u003c/li\u003e\n\u003cli\u003eIkilezi G, Augusto OJ, Sbarra A, Sherr K, Dieleman JL, Lim SS. Determinants of geographical inequalities for DTP3 vaccine coverage in sub-Saharan Africa. Vaccine. 2020 Apr 16;38(18):3447\u0026ndash;54. doi:10.1016/j.vaccine.2020.03.005 PubMed PMID: 32204938.\u003c/li\u003e\n\u003cli\u003eNilsen K, Tejedor-Garavito N, Leasure DR, Utazi CE, Ruktanonchai CW, Wigley AS, et al. A review of geospatial methods for population estimation and their use in constructing reproductive, maternal, newborn, child and adolescent health service indicators. BMC Health Serv Res. 2021 Sep 1;21. doi:10.1186/s12913-021-06370-y PubMed PMID: 34511089.\u003c/li\u003e\n\u003cli\u003eIlesanmi MM, Adeyinka DA, Olakunde BO. Sustainable Development Goals and childhood measles vaccination in Ekiti State, Nigeria: Results from spatial and interrupted time series analyses. Vaccine. 2022 Jun 21;40(28):3861\u0026ndash;8. doi:10.1016/j.vaccine.2022.05.037 PubMed PMID: 35644673.\u003c/li\u003e\n\u003cli\u003eSingh R, Agarwal S, De Figueiredo A, Mishra M, Agarwal D. Quantifying the role of maternal recall in estimates of routine immunisation rates in India: a large-scale sub-national Bayesian modelling study. In: COMPASS 2024 - Proceedings of the ACM SIGCAS/SIGCHI Conference on Computing and Sustainable Societies. Association for Computing Machinery, Inc; 2024. p. 74\u0026ndash;84. doi:10.1145/3674829.3675062\u003c/li\u003e\n\u003cli\u003eGamachu M, Mussa I, Deressa A, Tolera M, Birhanu A, Getachew T, et al. Patterns of basic pneumococcal conjugated vaccine coverage in Ethiopia from 2015 to 2018; further analysis of Ethiopian DHS (2016\u0026ndash;2019). Vaccine: X. Elsevier Ltd; 2024. doi:10.1016/j.jvacx.2024.100428\u003c/li\u003e\n\u003cli\u003eTesfa GA, Yehualashet DE, Getnet A, Bimer KB, Seboka BT. Spatial distribution of complete basic childhood vaccination and associated factors among children aged 12\u0026ndash;23 months in Ethiopia. A spatial and multilevel analysis. PLoS One. 2023 Jan 1;18(1 January). doi:10.1371/journal.pone.0279399 PubMed PMID: 36608032.\u003c/li\u003e\n\u003cli\u003eArtur Manuel Muloliwa, Luiz Antonio Bastos Camacho, Jos\u0026eacute; Fernando Souza Verani, Tayn\u0026atilde;na C\u0026eacute;sar Sim\u0026otilde;es, Martinho do Carmo Dgedge. Impact of vaccination on the incidence of measles in Mozambique in the period 2000 to 2011. Cad Saude Publica. 2013. doi:10.1590/s0102-311x2013000200013\u003c/li\u003e\n\u003cli\u003eAgimas MC, Asmamaw M, Mekonen N, Mamo F, Shewaye DA. Spatial pattern and associated factors of timely vaccination in Ethiopia using EDHS-2016 data: A multilevel and spatial analysis. PLoS One. 2024 Feb 1;19(2 February). doi:10.1371/journal.pone.0296123 PubMed PMID: 38324535.\u003c/li\u003e\n\u003cli\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 Sep 5;20(1). doi:10.1186/s12889-020-09461-3 PubMed PMID: 32891120.\u003c/li\u003e\n\u003cli\u003eAwol M, Alemu ZA, Moges NA, Jemal K. Geographical variations and associated factors of defaulting from immunization among children aged 12 to 23 months in Ethiopia: using spatial and multilevel analysis of 2016 Ethiopian Demographic and Health Survey. Environ Health Prev Med. 2021 Dec 1;26(1). doi:10.1186/s12199-021-00984-8 PubMed PMID: 34118886.\u003c/li\u003e\n\u003cli\u003eBantie GM, Tadege M, Nigussie TZ, Woya AA, Tekile AK, Melese AA, et al. Regional disparities of full pentavalent vaccine uptake and the determinants in Ethiopia: Mapping and spatial analysis using the EDHS data. PLoS One. 2025 Jan 1;20(1 January). doi:10.1371/journal.pone.0312514 PubMed PMID: 39787194.\u003c/li\u003e\n\u003cli\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 Vaccin Immunother. 2022;18(1). doi:10.1080/21645515.2022.2035558 PubMed PMID: 35148252.\u003c/li\u003e\n\u003cli\u003eShiferie F, Gebremedhin S, Andargie G, Tsegaye DA, Alemayehu WA, Fenta TG. Low Measles Vaccination Coverage and Spatial Analysis of High Measles Vaccination Dropout in Ethiopia\u0026rsquo;s Underprivileged Areas. Vaccines (Basel). 2024 Mar 1;12(3). doi:10.3390/vaccines12030328\u003c/li\u003e\n\u003cli\u003eAgegnehu CD, Alem AZ. Exploring spatial variation in BCG vaccination among children 0-35 months in Ethiopia: Spatial analysis of Ethiopian Demographic and Health Survey 2016. BMJ Open. BMJ Publishing Group; 2021. doi:10.1136/bmjopen-2020-043565 PubMed PMID: 33910946.\u003c/li\u003e\n\u003cli\u003eHailegebireal AH, Hailegebreal S, Tirore LL, Wolde BB. Spatial variation and predictors of incomplete pneumococcal conjugate vaccine (PCV) uptake among children aged 12\u0026ndash;35 months in Ethiopia: spatial and multilevel analyses. Front Public Health. 2024;12. doi:10.3389/fpubh.2024.1344089 PubMed PMID: 38864011.\u003c/li\u003e\n\u003cli\u003eAkinyemi YC. Geospatial Analysis of Vaccination Coverage and Predictors of Non- and Under-Vaccination of Children Aged 12\u0026ndash;23 Months in Nigeria. Papers in Applied Geography. 2021;7(1):1\u0026ndash;33. doi:10.1080/23754931.2020.1809009\u003c/li\u003e\n\u003cli\u003eKhan J, Shil A, Prakash R. Exploring the spatial heterogeneity in different doses of vaccination coverage in India. PLoS One. 2018 Nov 1;13(11). doi:10.1371/journal.pone.0207209 PubMed PMID: 30485291.\u003c/li\u003e\n\u003cli\u003eDhalaria P, Kumar P, Verma A, Priyadarshini P, Kumar Singh A, Tripathi B, et al. Exploring landscape of measles vaccination coverage: A step towards measles elimination goal in India. Vaccine. 2024 Jun 20;42(17):3637\u0026ndash;46. doi:10.1016/j.vaccine.2024.04.075 PubMed PMID: 38704248.\u003c/li\u003e\n\u003cli\u003eTamir TT, Terefe B, Wassie M, Workneh BS, Zegeye AF. Prevalence, spatial distribution and determinants of complete childhood pneumococcal conjugate vaccination in Ethiopia: spatial and multilevel analyses. BMC Public Health. 2024 Dec 1;24(1). doi:10.1186/s12889-024-19926-4 PubMed PMID: 39243021.\u003c/li\u003e\n\u003cli\u003eKamal A, Waseem A, Siddiqa M, Ijaz M, Shakeel A, Iftikhar S. Contextual factors influencing incomplete immunization and investigation of its geospatial heterogeneity in Pakistan: a cross-sectional study based on PDHS (2017\u0026ndash;18). BMC Public Health. 2023 Dec 1;23(1). doi:10.1186/s12889-023-16508-8 PubMed PMID: 37620868.\u003c/li\u003e\n\u003cli\u003eForzy T, Tesfaye L, Getnet F, Misganew A, Lamma SW, Worku A, et al. Assessing the geographic and socioeconomic determinants of vaccine coverage in Ethiopia: A spatial and multistage analysis at the district level. Vaccine. 2025 Apr 19;53. doi:10.1016/j.vaccine.2025.126834 PubMed PMID: 40056895.\u003c/li\u003e\n\u003cli\u003eDiress F, Negesse Y, Worede DT, Bekele Ketema D, Geitaneh W, Temesgen H. Multilevel and geographically weighted regression analysis of factors associated with full immunization among children aged 12\u0026ndash;23 months in Ethiopia. Sci Rep. 2024 Dec 1;14(1). doi:10.1038/s41598-024-74189-8 PubMed PMID: 39349770.\u003c/li\u003e\n\u003cli\u003eTamir TT, Workneh BS, Mekonen EG, Zegeye AF. Coverage, spatial distribution and determinants of childhood inactivated poliovirus vaccine immunization in Ethiopia. PLoS One. 2024 May 1;19(5 May). doi:10.1371/journal.pone.0301933 PubMed PMID: 38820454.\u003c/li\u003e\n\u003cli\u003eZegeye AF, Mekonen EG, Tekeba B, Alemu TG, Ali MS, Gonete AT, et al. Spatial distribution and determinants of measles vaccination dropout among under-five children in Ethiopia: A spatial and multilevel analysis of 2019 Ethiopian demographic and health survey. PLoS One. 2024 Jul 1;19(7 July). doi:10.1371/journal.pone.0305393 PubMed PMID: 38976660.\u003c/li\u003e\n\u003cli\u003eDefar A, Okwaraji YB, Tigabu Z, Persson L\u0026Aring;, Alemu K. Geographic differences in maternal and child health care utilization in four Ethiopian regions; A cross-sectional study. Int J Equity Health. 2019 Nov 12;18(1). doi:10.1186/s12939-019-1079-y PubMed PMID: 31718658.\u003c/li\u003e\n\u003cli\u003eAgimas MC, Belew AK, Sisay M, Baffa LD, Gashaw M, Abriham ZY, et al. Spatial variations and determinants of timely completion of vaccination in Ethiopia using further analysis of EDHS 2019 data: Spatial and multilevel analysis. PLoS One. 2024 Apr 1;19(4 April). doi:10.1371/journal.pone.0301409 PubMed PMID: 38578785.\u003c/li\u003e\n\u003cli\u003eRoot ED, Lucero M, Nohynek H, Anthamatten P, Thomas DSK, Tallo V, et al. Distance to health services affects local-level vaccine efficacy for pneumococcal conjugate vaccine (PCV) among rural Filipino children. Proc Natl Acad Sci U S A. 2014 Mar 4;111(9):3520\u0026ndash;5. doi:10.1073/pnas.1313748111 PubMed PMID: 24550454.\u003c/li\u003e\n\u003cli\u003ePuri P, Khan J, Shil A, Ali M. A cross-sectional study on selected child health outcomes in India: Quantifying the spatial variations and identification of the parental risk factors. Sci Rep. 2020 Dec 1;10(1). doi:10.1038/s41598-020-63210-5 PubMed PMID: 32313048.\u003c/li\u003e\n\u003cli\u003eTeshale AB, Amare T. Exploring spatial variations and the individual and contextual factors of uptake of measles-containing second dose vaccine among children aged 24 to 35 months in Ethiopia. PLoS One. 2023 Jan 1;18(1 January). doi:10.1371/journal.pone.0280083 PubMed PMID: 36598928.\u003c/li\u003e\n\u003cli\u003eKhan J, Shil A, Puri P. Population and sub-national (district) level diversity in missed and dropout of different doses of hepatitis-B vaccine among Indian children aged 12\u0026ndash;59 months. PLOS Global Public Health. 2022 May 17;2(5). doi:10.1371/journal.pgph.0000243\u003c/li\u003e\n\u003cli\u003eKumar G, R.S. R. Availability of public health facilities and utilization of maternal and child health services in districts of India. Clin Epidemiol Glob Health. 2022 May 1;15. doi:10.1016/j.cegh.2022.101070\u003c/li\u003e\n\u003cli\u003ePanda BK, Kumar G, Mishra S. Understanding the full-immunization gap in districts of India: A geospatial approach. Clin Epidemiol Glob Health. 2020 Jun 1;8(2):536\u0026ndash;43. doi:10.1016/j.cegh.2019.11.010\u003c/li\u003e\n\u003cli\u003eTegegne KT, Tegegne ET, Tessema MK, Wudu TK, Chalachew AA, Workneh AZ. Spatial Distribution and Associated Factors of Measles Vaccination Among Children Aged 12\u0026ndash;23 Months in Mozambique 2022\u0026ndash;2023: A Spatial and Multilevel Analysis. Bratislava Medical Journal. 2025 Feb 1;126(2):181\u0026ndash;95. doi:10.1007/s44411-024-00008-1\u003c/li\u003e\n\u003cli\u003eDas T. Unveiling Spatial Heterogeneity: A Study of Diverse Child Immunization Indicators Across Indian Districts. Child Indic Res. 2024 Aug 1;17(4):1615\u0026ndash;59. doi:10.1007/s12187-024-10135-4\u003c/li\u003e\n\u003cli\u003eSingh SK, Vishwakarma D. Spatial heterogeneity in the coverage of full immunization among children in India: Exploring the contribution of immunization card. Child Youth Serv Rev. 2021 Feb 1;121. doi:10.1016/j.childyouth.2020.105701\u003c/li\u003e\n\u003cli\u003eGeremew TT, Gezie LD, Abejie AN. Geographical variation and associated factors of childhood measles vaccination in Ethiopia: A spatial and multilevel analysis. BMC Public Health. 2019 Aug 30;19(1). doi:10.1186/s12889-019-7529-z PubMed PMID: 31470822.\u003c/li\u003e\n\u003cli\u003eChauhadry IA, Soofi SB, Sajid M, Ali R, Khan A, Naqvi SK, et al. Bridging the Vaccination Equity Gap: A Community-Driven Approach to Reduce Vaccine Inequities in Polio High-Risk Areas of Pakistan. Vaccines (Basel). 2024 Dec 1;12(12). doi:10.3390/vaccines12121340\u003c/li\u003e\n\u003cli\u003eKhan J, Shil A, Mohanty SK. Hepatitis B vaccination coverage across India: Exploring the spatial heterogeneity and contextual determinants. BMC Public Health. 2019 Sep 12;19(1). doi:10.1186/s12889-019-7534-2 PubMed PMID: 31510967.\u003c/li\u003e\n\u003cli\u003eUthman OA, Adedokun ST, Olukade T, Watson S, Adetokunboh O, Adeniran A, et al. Children who have received no routine polio vaccines in Nigeria: Who are they and where do they live? Hum Vaccin Immunother. 2017 Sep 2;13(9):2111\u0026ndash;22. doi:10.1080/21645515.2017.1336590 PubMed PMID: 28665749.\u003c/li\u003e\n\u003cli\u003eBrownwright TK, Dodson ZM, Van Panhuis WG. Spatial clustering of measles vaccination coverage among children in sub-Saharan Africa. BMC Public Health. 2017 Dec 15;17(1). doi:10.1186/s12889-017-4961-9 PubMed PMID: 29246217.\u003c/li\u003e\n\u003cli\u003eBantie B, Atnafu Gebeyehu N, Adella GA, Ambaw Kassie G, Mengstie MA, Abebe EC, et al. Mapping geographical inequalities of incomplete immunization in Ethiopia: a spatial with multilevel analysis. Front Public Health. 2024;12. doi:10.3389/fpubh.2024.1339539 PubMed PMID: 38912271.\u003c/li\u003e\n\u003cli\u003eDimitrova A, Carrasco-Escobar G, Richardson R, Benmarhnia T. Essential childhood immunization in 43 low- and middle-income countries: Analysis of spatial trends and socioeconomic inequalities in vaccine coverage. PLoS Med. 2023 Jan 1;20(1). doi:10.1371/journal.pmed.1004166 PubMed PMID: 36649359.\u003c/li\u003e\n\u003cli\u003eRerolle F, Dey AK, Benmarhnia T, Arnold BF. Spatial targeting and integration across vaccination, vitamin A and deworming programs throughout India 2019\u0026ndash;21. Int J Epidemiol. 2024 Dec 1;53(6). doi:10.1093/ije/dyae160 PubMed PMID: 39657980.\u003c/li\u003e\n\u003cli\u003eUmer MF, Zofeen S, Hu W, Qi X, Zhuang G. Spatiotemporal clustering analysis of Expanded Program on Immunization (EPI) vaccination coverage in Pakistan. Sci Rep. 2020 Dec 1;10(1). doi:10.1038/s41598-020-67839-0 PubMed PMID: 32620798.\u003c/li\u003e\n\u003cli\u003eKetsela A, Gebreyesus SH, Deressa W. Spatial distribution of under immunization among children 12\u0026ndash;23 months old in Butajira HDSS, southern Ethiopia. BMC Pediatr. 2021 Dec 1;21(1). doi:10.1186/s12887-021-02690-4 PubMed PMID: 33971837.\u003c/li\u003e\n\u003cli\u003eUtazi CE, Nilsen K, Pannell O, Dotse-Gborgbortsi W, Tatem AJ. District-level estimation of vaccination coverage: Discrete vs continuous spatial models. Stat Med. 2021 Apr 1;40(9):2197\u0026ndash;211. doi:10.1002/sim.8897 PubMed PMID: 33540473.\u003c/li\u003e\n\u003cli\u003eAlegana VA, Atkinson PM, Pezzulo C, Sorichetta A, Weiss D, Bird T, et al. Fine resolution mapping of population age-structures for health and development applications. J R Soc Interface. 2015 Apr 6;12(105). doi:10.1098/rsif.2015.0073 PubMed PMID: 25788540.\u003c/li\u003e\n\u003cli\u003eColson KE, Dwyer-Lindgren L, Achoki T, Fullman N, Schneider M, Mulenga P, et al. Benchmarking health system performance across districts in Zambia: A systematic analysis of levels and trends in key maternal and child health interventions from 1990 to 2010. BMC Med. 2015 Apr 2;13(1). doi:10.1186/s12916-015-0308-5 PubMed PMID: 25889124.\u003c/li\u003e\n\u003cli\u003ePramanik S, Muthusamy N, Gera R, Laxminarayan R. Vaccination coverage in India: A small area estimation approach. Vaccine. 2015 Mar 30;33(14):1731\u0026ndash;8. doi:10.1016/j.vaccine.2015.01.083 PubMed PMID: 25681661.\u003c/li\u003e\n\u003cli\u003eRoberts DA, Ng M, Ikilezi G, Gasasira A, Dwyer-Lindgren L, Fullman N, et al. Benchmarking health system performance across regions in Uganda: A systematic analysis of levels and trends in key maternal and child health interventions, 1990-2011. BMC Med. 2015 Dec 3;13(1). doi:10.1186/s12916-015-0518-x PubMed PMID: 26631048.\u003c/li\u003e\n\u003cli\u003eWollum A, Burstein R, Fullman N, Dwyer-Lindgren L, Gakidou E. Benchmarking health system performance across states in Nigeria: A systematic analysis of levels and trends in key maternal and child health interventions and outcomes, 2000-2013. BMC Med. 2015 Sep 2;13(1). doi:10.1186/s12916-015-0438-9 PubMed PMID: 26329607.\u003c/li\u003e\n\u003cli\u003eDong TQ, Wakefield J. SPACE-TIME SMOOTHING MODELS FOR SUBNATIONAL MEASLES ROUTINE IMMUNIZATION COVERAGE ESTIMATION WITH COMPLEX SURVEY DATA. Annals of Applied Statistics. 2021 Dec 1;15(4):1959\u0026ndash;79. doi:10.1214/21-AOAS1474\u003c/li\u003e\n\u003cli\u003eMercer LD, Safdar RM, Ahmed J, Mahamud A, Khan MM, Gerber S, et al. Spatial model for risk prediction and sub-national prioritization to aid poliovirus eradication in Pakistan. BMC Med. 2017 Oct 11;15(1). doi:10.1186/s12916-017-0941-2 PubMed PMID: 29017491.\u003c/li\u003e\n\u003cli\u003eSbarra AN, Rolfe S, Nguyen JQ, Earl L, Galles NC, Marks A, et al. Mapping routine measles vaccination in low- and middle-income countries. Nature. 2021 Jan 21;589(7842):415\u0026ndash;9. doi:10.1038/s41586-020-03043-4\u003c/li\u003e\n\u003cli\u003eAtalell KA, Alemayehu MA, Teshager NW, Belay GM, Alemu TG, Anlay DZ, et al. Mapping BCG vaccination coverage in Ethiopia between 2000 and 2019. BMC Infect Dis. 2022 Dec 1;22(1). doi:10.1186/s12879-022-07547-4 PubMed PMID: 35739462.\u003c/li\u003e\n\u003cli\u003eUtazi CE, Thorley J, Alegana VA, Ferrari MJ, Takahashi S, Metcalf CJE, et al. Mapping vaccination coverage to explore the effects of delivery mechanisms and inform vaccination strategies. Nat Commun. 2019 Dec 1;10(1). doi:10.1038/s41467-019-09611-1 PubMed PMID: 30967543.\u003c/li\u003e\n\u003cli\u003eWariri O, Utazi CE, Okomo U, Metcalf CJE, Sogur M, Fofana S, et al. Mapping the timeliness of routine childhood vaccination in The Gambia: A spatial modelling study. Vaccine. 2023 Sep 7;41(39):5696\u0026ndash;705. doi:10.1016/j.vaccine.2023.08.004 PubMed PMID: 37563051.\u003c/li\u003e\n\u003cli\u003eAtalell KA, Liyew AM, Alene KA. Spatial distribution of rotavirus immunization coverage in Ethiopia: a geospatial analysis using the Bayesian approach. BMC Infect Dis. 2022 Dec 1;22(1). doi:10.1186/s12879-022-07825-1 PubMed PMID: 36352357.\u003c/li\u003e\n\u003cli\u003eSbarra AN, Rolfe S, Haeuser E, Nguyen JQ, Adamu A, Adeyinka D, et al. Estimating vaccine coverage in conflict settings using geospatial methods: a case study in Borno state, Nigeria. Sci Rep. 2023 Dec 1;13(1). doi:10.1038/s41598-023-37947-8 PubMed PMID: 37422502.\u003c/li\u003e\n\u003cli\u003eWolde HF, Clements ACA, Gilmour B, Alene KA. Spatial co-distribution of tuberculosis prevalence and low BCG vaccination coverage in Ethiopia. Sci Rep. 2024 Dec 1;14(1). doi:10.1038/s41598-024-68549-7 PubMed PMID: 39738221.\u003c/li\u003e\n\u003cli\u003eUtazi CE, Wagai J, Pannell O, Cutts FT, Rhoda DA, Ferrari MJ, et al. Geospatial variation in measles vaccine coverage through routine and campaign strategies in Nigeria: Analysis of recent household surveys. Vaccine. 2020 Mar 23;38(14):3062\u0026ndash;71. doi:10.1016/j.vaccine.2020.02.070 PubMed PMID: 32122718.\u003c/li\u003e\n\u003cli\u003eUtazi CE, Olowe ID, Chan HMT, Dotse-Gborgbortsi W, Wagai J, Umar JA, et al. Geospatial Variation in Vaccination Coverage and Zero-Dose Prevalence at the District, Ward and Health Facility Levels Before and After a Measles Vaccination Campaign in Nigeria. Vaccines (Basel). 2024 Dec 1;12(12). doi:10.3390/vaccines12121299\u003c/li\u003e\n\u003cli\u003eFerreira LZ, Utazi CE, Huicho L, Nilsen K, Hartwig FP, Tatem AJ, et al. Geographic inequalities in health intervention coverage \u0026ndash; mapping the composite coverage index in Peru using geospatial modelling. BMC Public Health. 2022 Dec 1;22(1). doi:10.1186/s12889-022-14371-7 PubMed PMID: 36397019.\u003c/li\u003e\n\u003cli\u003eDong TQ, Wakefield J. Modeling and presentation of vaccination coverage estimates using data from household surveys. Vaccine. 2021 Apr 28;39(18):2584\u0026ndash;94. doi:10.1016/j.vaccine.2021.03.007 PubMed PMID: 33824039.\u003c/li\u003e\n\u003cli\u003eTakahashi S, Metcalf CJE, Ferrari MJ, Tatem AJ, Lessler J. The geography of measles vaccination in the African Great Lakes region. Nat Commun. 2017 May 25;8. doi:10.1038/ncomms15585 PubMed PMID: 28541287.\u003c/li\u003e\n\u003cli\u003eKawakatsu Y, Mosser JF, Adolph C, Baffoe P, Cheshi F, Aiga H, et al. High-resolution mapping of essential maternal and child health service coverage in Nigeria: a machine learning approach. BMJ Open. 2024 Jun 10;14(6). doi:10.1136/bmjopen-2023-080135 PubMed PMID: 38858137.\u003c/li\u003e\n\u003cli\u003eUtazi CE, Thorley J, Alegana VA, Ferrari MJ, Nilsen K, Takahashi S, et al. A spatial regression model for the disaggregation of areal unit based data to high-resolution grids with application to vaccination coverage mapping. Stat Methods Med Res. 2019 Nov 1;28(10\u0026ndash;11):3226\u0026ndash;41. doi:10.1177/0962280218797362 PubMed PMID: 30229698.\u003c/li\u003e\n\u003cli\u003eTesfaye L, Forzy T, Getnet F, Misganaw A, Woldekidan MA, Wolde AA, et al. Estimating immunization coverage at the district level: A case study of measles and diphtheria-pertussis-tetanus-Hib-HepB vaccines in Ethiopia. PLOS Global Public Health. 2024 Jul 25;4(7). doi:10.1371/journal.pgph.0003404\u003c/li\u003e\n\u003cli\u003eR: The R Project for Statistical Computing [Internet]. [cited 2026 Mar 18]. Available from: https://www.r-project.org/\u003c/li\u003e\n\u003cli\u003eStatistical software for data science | Stata [Internet]. [cited 2026 Mar 18]. Available from: https://www.stata.com/\u003c/li\u003e\n\u003cli\u003eSpatial without Compromise \u0026middot; QGIS Web Site [Internet]. [cited 2025 Nov 28]. Available from: https://qgis.org/\u003c/li\u003e\n\u003cli\u003eImmunization | UNICEF [Internet]. [cited 2025 Nov 24]. Available from: https://www.unicef.org/immunization\u003c/li\u003e\n\u003cli\u003eGAVI. Gavi 2024 Annual Progress Report [Internet]. 2025. Available from: www.gavi.org/funding/financial-reports\u003c/li\u003e\n\u003cli\u003eHomepage | Institute for Health Metrics and Evaluation [Internet]. [cited 2025 Nov 28]. Available from: https://www.healthdata.org/\u003c/li\u003e\n\u003cli\u003eNigeria VaxPop II - WorldPop [Internet]. [cited 2025 Nov 28]. Available from: https://www.worldpop.org/current-projects/completed-projects/nigeria-vaxpop-ii/\u003c/li\u003e\n\u003cli\u003eNgo-Bebe D, Mechael P, Kwilu FN, Bukele TK, Langwana F, Lobukulu GL, et al. Assessing the use of geospatial data for immunization program implementation and associated effects on coverage and equity in the Democratic Republic of Congo. BMC Public Health. 2025 Dec 1;25(1). doi:10.1186/s12889-025-21578-x PubMed PMID: 39856652.\u003c/li\u003e\n\u003cli\u003eWonodi C, Farrenkopf BA. Defining the Zero Dose Child: A Comparative Analysis of Two Approaches and Their Impact on Assessing the Zero Dose Burden and Vulnerability Profiles across 82 Low- and Middle-Income Countries. Vaccines (Basel). 2023 Oct 1;11(10). doi:10.3390/vaccines11101543\u003c/li\u003e\n\u003cli\u003eHogan D, Gupta A. Why Reaching Zero-Dose Children Holds the Key to Achieving the Sustainable Development Goals. Vaccines (Basel). 2023;11(4). doi:10.3390/vaccines11040781\u003c/li\u003e\n\u003cli\u003eWonodi C, Farrenkopf BA. Defining the Zero Dose Child: A Comparative Analysis of Two Approaches and Their Impact on Assessing the Zero Dose Burden and Vulnerability Profiles across 82 Low- and Middle-Income Countries. Vaccines (Basel). 2023 Oct 1;11(10). doi:10.3390/vaccines11101543\u003c/li\u003e\n\u003cli\u003eThe DHS Program - Quality information to plan, monitor and improve population, health, and nutrition programs [Internet]. [cited 2025 Nov 28]. Available from: https://www.dhsprogram.com/\u003c/li\u003e\n\u003cli\u003eHome - DHIS2 [Internet]. [cited 2025 Nov 28]. Available from: https://dhis2.org/\u003c/li\u003e\n\u003cli\u003eRozman LM, Sartori AMC, Souza AA, Soarez PC de. Discrepancy between official records of the Brazilian National Immunization Program Information System and the household survey about vaccination coverage in Cubat\u0026atilde;o, S\u0026atilde;o Paulo: cross-sectional study, 2023. Epidemiologia e Servi\u0026ccedil;os de Sa\u0026uacute;de. 2025;34. doi:10.1590/s2237-96222025v34e20240523.en PubMed PMID: 40767724.\u003c/li\u003e\n\u003cli\u003eArambepola R, Keddie SH, Collins EL, Twohig KA, Amratia P, Bertozzi-Villa A, et al. Spatiotemporal mapping of malaria prevalence in Madagascar using routine surveillance and health survey data. Sci Rep. 2020 Dec 1;10(1). doi:10.1038/s41598-020-75189-0 PubMed PMID: 33093622.\u003c/li\u003e\n\u003cli\u003eMolenaar J, Kikula A, van Olmen J, Pembe A, Beňov\u0026aacute; L. Getting the numbers right: Power, creativity and \u0026lsquo;good\u0026rsquo; routine maternal and neonatal health data in Southern Tanzania. Soc Sci Med. 2025 Feb 1;366. doi:10.1016/j.socscimed.2024.117668 PubMed PMID: 39787826.\u003c/li\u003e\n\u003cli\u003eBharti N, Djibo A, Tatem AJ, Grenfell BT, Ferrari MJ. Measuring populations to improve vaccination coverage. Sci Rep. 2016;5. doi:10.1038/srep34541\u003c/li\u003e\n\u003cli\u003eBeaulieu A, Ducharme J, Thibeault C, Akani BC, Ziegler D, Hogan D, et al. The Immunization Agenda 2030 Strategy to reach zero-dose children in low-income and middle-income countries: a scoping review. BMJ Glob Health. 2025 Aug 5;10(8):e018293. doi:10.1136/BMJGH-2024-018293 PubMed PMID: 40763981.\u003c/li\u003e\n\u003cli\u003eDiggle PJ, Tawn JA, Moyeed RA. Model-Based Geostatistics. J R Stat Soc Ser C Appl Stat. 1998 Sep 1;47(3):299\u0026ndash;350. doi:10.1111/1467-9876.00113\u003c/li\u003e\n\u003cli\u003eGiorgi E, Fronterr\u0026egrave; C, Macharia PM, Alegana VA, Snow RW, Diggle PJ. Model building and assessment of the impact of covariates for disease prevalence mapping in low-resource settings: to explain and to predict. J R Soc Interface. 2021 Jun 2;18(179):20210104. doi:10.1098/rsif.2021.0104\u003c/li\u003e\n\u003cli\u003eGAVI. IMPROVING IMMUNISATION COVERAGE AND EQUITY THROUGH THE EFFECTIVE USE OF GEOSPATIAL TECHNOLOGIES Improving Immunisation Coverage and Equity through the Effective Use of Geospatial Technologies and Data [Internet]. 2020 [cited 2025 Nov 28]. Available from: https://www.gavi.org/our-impact/evaluation-studies/using-geospatial-technologies-improve-immunisation-coverage-equity\u003c/li\u003e\n\u003cli\u003eGiorgi E, Macharia PM, Woodmansey J, Snow RW, Rowlingson B. Maplaria: a user friendly web-application for spatio-temporal malaria prevalence mapping. Malaria Journal 2021 20:1. 2021 Dec 20;20(1):471-. doi:10.1186/S12936-021-04011-7 PubMed PMID: 34930265.\u003c/li\u003e\n\u003cli\u003eJohnson O, Fronterre C, Diggle PJ, Amoah B, Giorgi E. MBGapp: A Shiny application for teaching model-based geostatistics to population health scientists. PLoS One. 2021 Dec 1;16(12):e0262145. doi:10.1371/JOURNAL.PONE.0262145 PubMed PMID: 34972193.\u003c/li\u003e\n\u003cli\u003eWu Y, Dong Q, Xu J, Li ZR, Wakefield J. sae4health: An R Shiny Application for Small Area Estimation in Low- and Middle-Income Countries. ArXiv [Internet]. 2025 May 2 [cited 2025 Nov 28]. Available from: https://arxiv.org/pdf/2505.01467\u003c/li\u003e\n\u003cli\u003eEspinosa L, Salath\u0026eacute; M. Use of large language models as a scalable approach to understanding public health discourse. PLOS Digital Health. 2024 Oct 1;3(10):e0000631. doi:10.1371/JOURNAL.PDIG.0000631\u003c/li\u003e\n\u003cli\u003eBurstein R, Mafuta E, Proctor JL. Large language models for analyzing open text in global health surveys: why children are not accessing vaccine services in the Democratic Republic of the Congo. Int Health. 2025 Sep 1;17(5):843\u0026ndash;52. doi:10.1093/INTHEALTH/IHAF015 PubMed PMID: 40052518.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[{"identity":"e638a558-ec6b-45f9-ae10-e6300b65a185","identifier":"10.13039/501100001713","name":"European and Developing Countries Clinical Trials Partnership","awardNumber":"101159477","order_by":0},{"identity":"5b717e78-0a4a-4534-b08a-ab4270da888f","identifier":"10.13039/501100003130","name":"Fonds Wetenschappelijk Onderzoek","awardNumber":"1201925N","order_by":1}],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Institute of Tropical Medicine Antwerp","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":"zero-dose, immunization coverage, vaccination inequity, spatial, high-resolution, geostatistics, geospatial modelling, mapping, small-area estimation, data, Low- and lower-middle-income countries ","lastPublishedDoi":"10.21203/rs.3.rs-9304718/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9304718/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e Zero-dose (ZD) children remain a critical public health concern, particularly in low- and lower-middle-income countries (LLMICs), where over 80% of the global ZD population resides, disproportionately concentrated among the most marginalised. \u0026nbsp;Geospatial methods have emerged as effective tools for identifying and targeting immunization gaps. However, no review has systematically documented the spatial data and methods used to identify and characterize ZD children and corresponding gaps. This scoping review addresses that gap across LLMICs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e Following PRISMA-ScR guidelines, we searched for peer reviewed articles published upto 2025 on spatial modelling of childhood vaccination coverage in LLMICs using six databases: PubMed, Web of Science, Scopus, Cochrane, Embase, and EBSCOhost-CINAHL. We extracted details on study characteristics, covariate types and sources, modelling methods, and the gaps. Articles were thematically summarized focusing on geospatial data, modelling approaches, and their corresponding gaps.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003eOf 15,587 articles retrieved, 102 from 68 LLMICs were included, with 70% published between 2021 and 2024, and 87% concentrated in Ethiopia, Nigeria, and India. Only a fifth assessed ZD prevalence, based on two distinct definitions. Studies relied predominantly on household survey data, with routine administrative data underused. Covariate data were dominated by demographic factors (49%) with limited representation of hard-to-reach contexts. Methods included clustering and autocorrelation analysis (54%), spatial interpolation (45%), small-area estimation (13%), and machine learning (8%). Key gaps included inconsistent ZD definitions, missing covariates, data inaccuracies, sparse samples, and weak representation of conflict-affected, informal-settlement, and mobile populations.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions: \u003c/strong\u003eDespite growing availability of spatial data and methods, geospatial identification of ZD children remains concentrated in few countries, relies heavily on survey data, uses inconsistent definitions, and is constrained by limitations that systematically exclude the most marginalised populations. Addressing these gaps will require harmonised definitions, integrated data systems, and reproducible modelling approaches underpinned by sustained investment in local analytical capacity.\u003c/p\u003e","manuscriptTitle":"Geospatial approaches for mapping zero-dose children in low- and lower-1 middle-income countries: A scoping review","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-03 06:45:42","doi":"10.21203/rs.3.rs-9304718/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":"ae768e24-ac89-4050-966e-722af95e9a01","owner":[],"postedDate":"April 3rd, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":65626456,"name":"Epidemiology"},{"id":65626457,"name":"Statistical Epidemiology"},{"id":65626458,"name":"Biostatistics"},{"id":65626459,"name":"Medical Informatics"},{"id":65626460,"name":"Health Policy"},{"id":65626461,"name":"Geographic Information Systems"}],"tags":[],"updatedAt":"2026-04-03T06:45:42+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-03 06:45:42","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9304718","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9304718","identity":"rs-9304718","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}