Spatio-temporal patterns of Dengue cases and deaths, associated factors and identification of risk areas at two geographical scales: municipalities in Brazil and intra-urban areas of Campinas, State of São Paulo

Abstract

Dengue is a disease considered an important public health problem, whose incidence has expanded abruptly in recent decades. This increase has been linked, among other things, to changes in climatic conditions that can affect different aspects of the disease. Currently, there is no vaccine available and the identification of areas at risk for Dengue is given by entomological surveillance of immature forms of the vector, which has limitations and has been criticized. Therefore, it is important to consider and seek alternatives to control the disease. Objectives: to identify areas and risk factors for Dengue incidence, mortality and lethality, evaluate the degree of overlap in the results found and investigate to what extent the incidence could predict mortality and lethality, considering a space-time architecture and two geographical scales: Brazilian municipalities and intra-urban areas of Campinas, SP, in the period between 2000 and 2019. Methods: ecological study, in which the incidence, lethality and mortality of Dengue will be considered in all municipalities in Brazil and in the census sectors in the municipality of Campinas, thus having an inter- and intramunicipal view. For this, high-risk clusters (spatial, spatio-temporal and spatial variation of temporal trends) will be identified through scanning statistics, with the comparison of individual and multiple analysis, the latter considering the incidence and mortality together. Also, modeling of incidence, lethality and mortality will be made to identify areas of risk, considering demographic, socioeconomic, environmental, climatic and health care variables, using space-time architectures and Poisson probability distributions, in an approach considering Latent Gaussian Bayesians Models. Some intramunicipais variables will be obtained in conjunction with a larger project, by artificial and use of satellite images. The choice of model variables will be based on Directed Acyclic Graphs (DAGs) and, at the end, a single DAG will be built so that risk factors can be mapped. For Brazilian municipalities, lethality and mortality will also be predicted, based on the incidence prediction that is being returned in another project also linked to this one. (AU)