Spatial Distribution, Environmental Associations and Geographical Mapping of Microscopy-confirmed Malaria Cases among Children under Five in Nigeria

Abstract

Aims: This study assessed the prevalence and geographical distribution of microscopy-confirmed malaria among children aged 6–59 months in Nigeria, examined selected environmental associations and spatial clustering, modelled cluster-level malaria probability, and identified geographical areas with comparatively higher model-predicted probabilities. Study Design: Cross-sectional secondary analysis of nationally representative survey data integrated with environmental and geospatial datasets. Place and Duration of Study: Nigeria; analysis used data from the 2021 Nigeria Malaria Indicator Survey and environmental datasets corresponding to the survey period. Methodology: Data for 10,655 children from 567 survey clusters were analysed using survey-weighted prevalence estimation and complex-samples logistic regression. Cluster locations were linked with environmental data, and Local Moran’s I was used to identify statistically significant spatial clusters and outliers. Model-predicted probabilities were aggregated to cluster level and used to identify intervention-priority areas. Results: National malaria prevalence was 22.3%, ranging from 2.6% in Lagos to 49.2% in Kebbi. Prevalence was higher in rural than urban areas (26.7% versus 10.5%; P < .001). Annual rainfall was associated with malaria in one alternative model, while elevation and built-up land coverage showed inverse associations. Local Moran’s I identified 193 significant clusters and outliers, including 45 High–High hotspots and 107 Low–Low coldspots. Model-predicted probabilities of malaria across the survey clusters ranged from 1.7% to 37.3%; 272 of 567 clusters (48.0%) were classified as higher categories of model-predicted probability. Bauchi and Jigawa had the highest intervention-priority percentages (87.5% each), followed by Ebonyi (84.6%), Katsina (82.4%) and Kebbi (80.0%). Conclusion: Childhood malaria risk in Nigeria shows marked geographical variation. The identified hotspots and priority areas can support population-level surveillance, resource allocation and targeted malaria control. Given the model’s modest predictive performance and displacement of DHS cluster coordinates, the estimates should not be interpreted as individual-level clinical predictions.