Back

Spatial Distribution and Multilevel Determinants of Unvaccinated Children in Somaliland: An Analysis of the 2020 SDHS. Spatial and Multilevel Analysis

dahir, H. m.; Korse, A. H.; Nadarjah, s.

2025-09-18 public and global health
10.1101/2025.09.16.25335900 medRxiv
Show abstract

BackgroundUnvaccinated children, often termed zero-dose, represent a critical global public health challenge and a stark indicator of health system inequities, particularly in low- and middle-income countries. Somaliland faces an exceptionally severe crisis, with an alarmingly high prevalence of zero-dose children. This study aimed to comprehensively assess the spatial distribution and identify the individual and community-level determinants of unvaccinated children among those under five years in Somaliland, utilizing data from the 2020 Somaliland Demographic and Health Survey (SDHS). MethodsThis cross-sectional study analyzed data from 3,255 children under five from the 2020 SLHDS. Multilevel logistic regression analysis was employed to identify determinants of zero-dose status, comparing models using AIC, BIC, and Log Likelihood, and assessing community-level variation with the Intraclass Correlation Coefficient (ICC). Geospatial analyses, including descriptive mapping, Global Morans I, Local Morans I (LISA), and Getis-Ord Gi* statistics, were used to explore the spatial distribution and clustering of unvaccinated children. ResultsThe prevalence of unvaccinated children in Somaliland was found to be extremely high at 79.88%. The final multilevel model (Model III) indicated that significant individual-level factors associated with lower odds of being zero-dose included maternal secondary education (AOR=0.409), not currently breastfeeding (AOR=0.565), delivery in a health facility (AOR=0.642), belonging to middle (AOR=0.602) or rich (AOR=0.499) wealth quintiles, and attending any ANC visits (AORs 0.349 for <4 visits; 0.253 for [&ge;]4 visits). Spatial analysis revealed marked regional disparities, with Awdal (88.66%) and Sool (84.71%) exhibiting the highest zero-dose prevalence, while Sahil (71.43%) and Togdheer (75.33%) had the lowest rates. Global Morans I indicated overall negative spatial autocorrelation, suggesting a pattern of dissimilarity. Local Morans I identified clusters of lower zero-dose rates in Sahil, Maroodi Jeex, and Togdheer, while Awdal and Sool emerged as high-rate outliers. The initial ICC of 24.8% (Model 0) reduced to 7.5% (Model III), confirming substantial community-level influence on zero-dose status. ConclusionZero-dose vaccination status is alarmingly prevalent and exhibits significant spatial heterogeneity in Somaliland. Key determinants include maternal education, healthcare utilization (ANC visits, facility delivery), and household wealth. The geospatial patterns highlight specific high-risk regions and areas with clustering of lower rates, underscoring the need for geographically targeted and context-specific interventions to reach unreached children and improve immunization equity.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.