Modeling malaria incidence in Sucre state, Venezuela using a Bayesian approach

This paper presents a hierarchical Bayesian Poisson lognormal model for malaria incidence in Sucre state, Venezuela, during the period 1990−2002. The logarithm of the relative risk of the disease for each county or municipality is ex-pressed as an additive model that includes a multiple regression with social-economic and climatic covariates; a random effect that captures the spatial heterogeneity in the study region and a CAR (Conditionally Autoregressive) component, that recognizes the effect of nearby municipalities in the transmission of the disease each year. For most years the selected model captures well the spatial structure between the relative risks from the nearby municipalities. When a poor model fit is obtained, a t-Student model for the spatial heterogeneity parameter improves model fitting results. From the 15 municipalities in Sucre state during the study period 1990−2002, 7 of them presented high relative risks (greater than 1) in most years. These areas are mostly agricultural areas with poor living conditions.