Abstract
BACKGROUND: Malaria is one of the principal health problems in Mozambique, representing 48% of total external consultations and 63% of paediatric hospital admissions in rural and general hospitals with 26.7% of total mortality. Plasmodium falciparum is responsible for 90% of all infections being also the species associated with most severe cases. The aim of this study was to identify zones of high malaria risk, showing their spatially and temporal pattern. METHODS: Space and time Poison model for the analysis of malaria data is proposed. This model allows for the inclusion of environmental factors: rainfall, temperature and humidity as predictor variables. Modelling and inference use the fully Bayesian approach via Markov Chain Monte Carlo (MCMC) simulation techniques. The methodology is applied to analyse paediatric data arising from districts of Maputo province, Mozambique, between 2007 and 2008. RESULTS: Malaria incidence risk is greater for children in districts of Manhiça, Matola and Magude. Rainfall and humidity are significant predictors of malaria incidence. The risk increased with rainfall (relative risk-RR: .006761, 95% interval: .001874, .01304), and humidity (RR: .049, 95% interval: .03048, .06531). Malaria incidence was found to be independent of temperature. CONCLUSIONS: The model revealed a spatial and temporal pattern of malaria incidence. These patterns were found to exhibit a stable malaria transmission in most non-coastal districts. The findings may be useful for malaria control, planning and management.