Socioeconomic and community determinants of under-five mortality in the Dominican Republic: a multilevel analysis of national survey data

Background: Although the Dominican Republic has achieved considerable progress in under-five mortality (U5M) over the past decades, there still exists considerable socioeconomic and geographic inequality. Children from poorer households and disadvantaged communities continue to face disproportionately higher mortality risks. Nevertheless, research on the multilevel impact of individual- and community-level determinants on child survival in the country remains limited. We aimed to quantify the associations between individual- and community-level socioeconomic and demographic determinants and under-five mortality in the Dominican Republic, using a multilevel analytic approach applied to nationally representative survey data. Methods: We conducted a cross-sectional multilevel analysis using nationally representative birth history data from the 2013 Dominican Republic Demographic and Health Survey. We included approximately 13,100 children nested within 453 community clusters. We used multilevel logistic regression models with random effects to examine the relationship between socioeconomic, demographic, and community-based factors and U5M. We used the Akaike Information Criterion, Bayesian information criterion, and likelihood ratio tests to assess model fit. We quantified community-level variation using clustering measures such as the intraclass correlation coefficient and median odds ratio (OR). Results: Significant socioeconomic gradients in U5M were observed. Children in the poorest wealth quintile had nearly twice the odds of mortality compared to those in the richest quintile (adjusted OR (aOR) = 1.98; 95% confidence interval (CI) = 1.44-2.71). Maternal education was strongly associated with child survival, with children of mothers with no education having higher odds of mortality (aOR = 2.14; 95% CI = 1.58-2.89) compared to those with tertiary education. Unimproved sanitation (aOR = 1.38; 95% CI = 1.16-1.64), unimproved water sources (aOR = 1.27; 95% CI = 1.05-1.54), short birth intervals (<24 months) (aOR = 1.39; 95% CI = 1.18-1.64), and fewer than four antenatal care visits (aOR = 1.61; 95% CI = 1.32-1.96) were significant predictors. At the community level, rural residence (aOR = 1.51; 95% CI = 1.19-1.92), higher regional poverty (aOR = 1.43; 95% CI = 1.09-1.87), and lower healthcare access (aOR = 1.38; 95% CI = 1.05-1.82) were associated with increased mortality risk. Community-level variance decreased from 12.9% in the null model to 4.2% in the fully adjusted model, indicating meaningful contextual effects. Conclusions: Both individual socioeconomic conditions and broader community-level factors shape U5M in the Dominican Republic. Persistent inequalities in wealth, maternal education, and access to basic services highlight the need for integrated policy approaches targeting both household and structural determinants. Strengthening maternal education, expanding access to water and sanitation, and improving rural healthcare infrastructure are critical for reducing child mortality and advancing progress toward Sustainable Development Goal 3.2.