eISSN: 2378-315X
Submit manuscript...
Research Article Volume 15 Issue 1
Bayesian spatial modeling of malaria risk in Mozambique: a province-level analysis
Élio José Taero
Received: August 04, 2026 | Published: August 26, 2026
Citation: Taero EJ. Bayesian spatial modeling of malaria risk in Mozambique: a province-level analysis. Biom Biostat Int J. 2026;15(1):62-67. DOI: 10.15406/bbij.2026.15.00449
Abstract
Malaria remains one of the major public health challenges in Mozambique, with substantial spatial variation in disease occurrence across provinces. Understanding the geographical distribution of malaria risk is essential for guiding targeted control strategies and improving the allocation of health resources. This study investigated the spatial distribution of malaria relative risk across the provinces of Mozambique using a Bayesian spatial modelling approach. Global and local spatial autocorrelation analyses were performed to identify spatial patterns and clustering of malaria risk. The Intrinsic Conditional Autoregressive (ICAR) model was applied to estimate province-specific relative risks. Posterior exceedance probability maps were generated to assess the statistical evidence of increased or reduced malaria risk. The results revealed significant positive spatial autocorrelation, indicating that malaria risk is spatially structured rather than randomly distributed. A high-risk spatial cluster was identified in the northern and central regions of the country, particularly encompassing the provinces of Nampula, Cabo Delgado, and Zambézia, with Nampula emerging as the main hotspot. In contrast, most provinces in the southern region and some central provinces exhibited low relative risk, with the exception of Inhambane, which showed elevated relative risk. Although Niassa presented relatively low estimated risk, its location adjacent to high-risk provinces suggests increased epidemiological vulnerability. The ICAR model provided robust estimates of relative risk, while posterior exceedance probabilities confirmed areas with strong evidence of risk exceeding the expected level. The findings demonstrate substantial spatial heterogeneity in malaria risk across Mozambique and highlight the importance of incorporating spatial dependence into disease mapping analyses. Bayesian spatial models and posterior exceedance probability maps provide valuable evidence to support geographically targeted malaria interventions.
Keywords: malaria, Mozambique, bayesian disease mapping, ICAR model, relative risk, posterior exceedance probability
©2026 Taero. This is an open access article distributed under the terms of the, which permits unrestricted use, distribution, and build upon your work non-commercially.
Citations
Rejected Articles
2 7