Dengue em Belo Horizonte: distribuição espaço-temporal (1996 A 2019), predição de casos (2007 A 2020) e associação com variáveis socioambientais (2010)

Abstract

Dengue is a re-emerging arbovirus with wide endemic occurrence in the southern hemisphere. In the Americas the occurrences increase, in 40 years the number of accumulated cases increased from 1.5 million in the 1980s to 16.2 million in the 2010-2019 decade. The number of cases in the Americas in 2019 reaffirmed the seriousness, there were about 3.1 million notifications, with 28,000 serious cases and 1,534 deaths. This thesis reports studies on the twenty-four years of dengue in Belo Horizonte, MG, Brazil (1996 to 2020) with the following objectives: analyze the temporal behavior of cases in different periods, using predictive models (article 1) and investigate the association between dengue and social and environmental variables in 2010 (article 2) and know the spatiotemporal pattern of cases and the positivity of ovitraps (analysis exploratory in progress). Ecological approaches were adopted in the studies; the first followed the behavior of cases in the period 2007 to 2020, using the "Seasonal Integrated Autoregressive Moving Average (SARIMA) to predict the years from 2016 to 2020. In the second study, the association between the risk of occurrence of dengue and the Index was sought of Normalized Difference Vegetation (NDVI), until then not investigated in the city, in regression models, cross-sectional, multi-group analytical study. In the exploratory study, several spatial analyzes were used, the spatial dependence by Moran's I statistics and the Local Indicator of Spatial Association (LISA) for dengue and ovitrap cases, the Kernel Ratio statistic, and the scan statistic (Scan) for cases of dengue. The historical series of dengue in Belo Horizonte showed, after the 1996 outbreak, five epidemics in 1998, 2010, 2013, 2016 and 2019. The incidence rates were increasing (3,755/2,053/ 3,949/6,176 and 4,894.5 cases per 100,000 inhabitants) and resulting from high transmission. The SARIMA prediction model (3.1,1) (0.0,0)12 for the periods 2007 to 2020 indicated the expected occurrence of epidemic cases in 2016 and 2019 and non-epidemic cases in 2017, 2018 and 2020. However, the wide confidence interval of the models raises questions about their robustness. The presence of green in the environment can be related to the occurrence of dengue, modified by socioeconomic vulnerability (IVS). In the exploratory study, the local Moran statistic (LISA) indicated local spatial dependence, except in 2000. Northern residents in the city were almost 23 times more likely to have dengue compared to those in the south. Spatio-temporal scan statistics identified 22 significant clusters, with the relative risk ranging from 2.65 (2010) to 22.88 (2013), half of them in the northern region. The SARIMA model should be revised when predicting epidemiological patterns, considering the importance of looking for factors that interfere with transmission patterns, and expanding the investigation of prediction models in disease surveillance systems. Social and environmental characteristics related to the presence of vegetation seem to be associated, especially in vulnerable areas. Agglomeration and dispersion of dengue cases over time should guide work on host/virus/vector mobility and indicate priority regions for intervention. The studies carried out allowed analyzes to understand the pattern of occurrence of dengue and can guide public control policies.