Variáveis ambientais e climáticas associadas à incidência de dengue no estado de Mato Grosso, Brasil

Abstract

Dengue has been a major global challenge and a subject of studies that involves the most diverse areas of knowledge, due to its complexity, scale and difficulty of control. It is considered endemic in practically all tropical and subtropical regions and, today, it is the most prevalent and incident arbovirus in the world, constituting a serious public health problem. The epidemiology of dengue is greatly affected by the spatial distribution of its vector Aedes aegypti, and this, in turn, is influenced by environmental and climatic factors. The state of Mato Grosso, Brazil, located in the Legal Amazon, has an extensive territorial area, tropical climate, three biomes, cities in full urban expansion and a history of successive dengue epidemics. However, there are few time series studies that analyze the incidence of dengue in the state. This study aimed to analyze the influence of environmental and climatic variables on the incidence of dengue in the state of Mato Grosso, in the period from 2008 to 2020, and proposes a forecast model for the municipality of Sinop. The influence of environmental variables on the dengue incidence rate, accumulated over 11 years, in the 141 municipalities in the state of Mato Grosso was analyzed by the Generalized Linear Model. The influence of climatic variables on the monthly incidence rate of dengue in Sinop, MT, in the 13 years of study, was analyzed by the Generalized Linear Moving Averages Model (GLARMA). The results indicated that municipalities with a predominance of forest (Amazon and Pantanal), with higher percentages of deforestation over 31 years (1988-2018) showed higher rates of accumulated dengue incidence from 2008 to 2018. In Sinop, the higher the precipitation (t-2), maximum temperature (t-1) and relative humidity (t-1), the greater the incidence of dengue. The GLARMA model under study was able to capture the variations of the time series, with adjusted values very close to the real ones, which indicates the good accuracy of the model and, consequently, its efficiency in predicting the incidence of dengue in the following month.