Hepatitis B vaccination coverage and factors associated with higher uptake among children aged 5–35 months in eight selected sub-Saharan African countries: a multilevel ordinal logistic regression analysis

Sub-Saharan Africa (SSA) carries the world’s heaviest burden of chronic hepatitis B, most of which is acquired in early life; however, coverage remains short of what is needed to interrupt transmission. Most evidence treats childhood vaccination as a yes-or-no outcome and pools different antigens, hiding both the partial vaccination typical of a multi-dose schedule and the individual, household, and community influences that act together on uptake. We examined the factors associated with higher hepatitis B vaccination rates among children aged 5–35 months in eight selected SSA countries. We pooled the most recent Demographic and Health Survey (DHS) data (2016–2024) from the eight SSA countries whose surveys recorded all four hepatitis B touchpoints (birth dose and three pentavalent doses): 49,075 children aged 5–35 months (weighted n = 48,124) nested within 41,476 households and 5,580 communities. Hepatitis B vaccination was defined as an ordered outcome (not, partially or fully vaccinated) from the number of recorded doses. Guided by Andersen’s Behavioural Model of Health Services Use, we fitted three-level random-intercept ordinal logistic regression models (children within households within communities) with country fixed effects, reporting adjusted odds ratios (aOR) with 95% confidence intervals (CI) and intra-cluster correlation coefficients (ICC). Descriptive estimates accounted for the survey weights, strata and clusters; the proportional odds assumption was tested, and design-based, threshold-specific and country-stratified models were fitted as sensitivity analyses. Overall, 38.3% (95% CI 37.5–39.1) of children were fully vaccinated, 46.2% (45.4–46.9) partially vaccinated and 15.6% (14.9–16.2) not vaccinated, and only 49.7% (48.9–50.6) had a recorded birth dose. Full vaccination ranged from 3.4% in Burundi and 11.4% in Sierra Leone, where the birth dose was not part of the national schedule, to 89.7% in The Gambia; Nigeria had the largest share of unvaccinated children (35.3%). In the full model, facility-based delivery (aOR 2.83, 95% CI 2.66–3.02), maternal education (aOR 3.20, 95% CI 2.79–3.66 for higher versus none), exposure to media (aOR 1.29, 95% CI 1.22–1.36), maternal employment (aOR 1.27, 95% CI 1.20–1.34), older maternal age, older child age and greater household wealth (aOR 1.60, 95% CI 1.43–1.79 for richest versus poorest) were associated with a higher vaccination category, whereas rural residence (aOR 0.78, 95% CI 0.70–0.86) and a preceding birth interval shorter than 24 months (aOR 0.86, 95% CI 0.79–0.94) were associated with a lower category. Health insurance coverage was not associated with vaccination once country was accounted for (aOR 0.98, 95% CI 0.89–1.07). The community level accounted for 32% of the residual variance and households for a further 3%, and the multilevel model fitted substantially better than the single-level model (LR χ² = 5978.93, p < 0.001). In these eight countries most children were only partially vaccinated against hepatitis B and half had no recorded birth dose, with large differences between countries that mirror national birth-dose policy. Within countries, uptake was associated with facility delivery, maternal education, media exposure and household wealth, and it clustered strongly within communities. These cross-sectional associations suggest that introducing and delivering the birth dose through maternal and newborn care, reaching home births, and targeting less-advantaged mothers and lagging communities may help close the gap toward the 2030 elimination targets.