Developing an Original Multi-Level Model of Child Survival Determinants in Northern Nigeria

Child survival in northern Nigeria is shaped by individual, household, community, and health system factors that operate simultaneously and interact across levels, and developing an original multi-level analytical framework that captures these interactions provides a methodologically advanced contribution to Nigerian child health demography. This study developed an original multi-level model of child survival determinants in northern Nigeria, drawing on theory, data analysis, and model development using the pooled NDHS 2013, 2018, and 2021 datasets. A multi-level theory-building and empirical methodology was employed over 20 months. Theoretical analysis synthesised Mosley-Chen proximate determinants framework, Heer and Smith's demand model for children, and community-level health systems theories. Multi-level logistic regression models were specified with children at Level 1, households at Level 2, communities at Level 3, and states at Level 4, applied to 42,000 under-five observations from the six northern states. Intra-class correlation analysis confirmed that 22.4 percent of child survival variance was explained by community-level factors and 12.1 percent by state-level factors, justifying multi-level modelling over single-level approaches. The original Northern Nigeria Child Survival Multi-Level Model identifies maternal education (Level 2), community health worker density (Level 3), and health facility delivery infrastructure investment (Level 4) as the three highest-leverage determinants. Cross-level interaction effects between maternal education and community healthcare access were statistically significant, demonstrating complementary rather than additive effects. Expert review confirmed the model's original methodological contribution.

Keywords: child survival, multi-level model, northern Nigeria, proximate determinants, NDHS