Prevalence mapping of under-five malaria cases in Nigeria using small area estimation approach

Malaria remains a significant public health problem in sub-Saharan Africa, with Nigeria accounting for a high proportion of the global cases. Accurate mapping of under-five malaria cases is critical for targeted interventions and for achieving Sustainable Development Goal 3. Previous estimates of under-five malaria prevalence in Nigeria have often neglected to account for complex survey designs, potentially biasing results. This study addressed this gap by applying small-area estimation (SAE) techniques that explicitly account for survey design to estimate under-five malaria prevalence at national and subnational levels. Data from the 2021 Nigeria Malaria Indicator Survey (NMIS) were analyzed (􀜰 􀵌 10,717 under-five children) using three approaches: a direct weighted estimator, a smoothed direct (Fay-Herriot) model incorporating survey design, and a Bayesian spatial hierarchical model (BYM2 smoothed model), which were computed using integrated nested Laplace approximation (INLA). The performance of the models was compared using the deviance information criterion (DIC) and the marginal likelihood. The smoothed direct model outperformed the BYM2 smoothed model (DIC: 39.99 vs. 287.75; marginal likelihood: 􀵆11.51 vs. 􀵆158.48). National underfive malaria prevalence was approximately 34.23% 􁈺95% CrI [31.19,37.37]), with significant spatial disparities. Hotspots were concentrated in the Northwest (Kebbi, Zamfara, and Jigawa) and Northeast (Yobe and Bauchi) regions. These findings demonstrate that accounting for survey design reduces bias in prevalence mapping. Targeted, region-specific interventions in identified hotspots are urgently needed. This study provides policymakers andpublic health practitioners with methodologically robust evidence to optimize malaria control strategies and resource allocation in Nigeria.