KDD aplicado à análise espaço-temporal da Leishmaniose Visceral no estado do Pará entre os anos de 2007 e 2019

Abstract

Visceral Leishmaniasis (LVH) is a severe and neglected parasitic disease affecting various regions worldwide. In Brazil, the state of Pará is known for harboring municipalities with a high transmission risk, standing out as one of the disease epicenters in the Americas during the triennium 2017-2019. This study aimed to identify high-risk areas within the state of Pará and investigate their temporal and spatial distribution over the years. We employed a Knowledge Discovery in Databases (KDD) approach to conduct a retrospective exploratory analysis based on VL notification records from the Notifiable Diseases Information System (SINAN), covering the period from 2007 to 2019. The KDD process was adopted as the primary methodology for data analysis. The results revealed that the Global Moran's I index was consistently positive throughout the analyzed period, indicating the presence of spatial clusters, reaching its highest value in 2019 (0.45). The analysis of Local Moran's I indices (LISA statistics) highlighted municipalities with higher risk concentrated mainly in the northeastern and southeastern regions of the state. Furthermore, the application of Space-Time Scan Statistics (SaTScan) identified three distinct circular clusters, with a radius of up to 178 km, two of which were observed between 2007 and 2013 in the northeastern region of Pará and another between 2014 and 2019 in the southeastern part of the state.Human Visceral Leishmaniasis (LVH) is a serious and neglected parasitic disease present in several continents. In Brazil, the Para state has the municipalities with the highest transmission risk in the Americas, considering the period 2017-2019. This study aimed to identify risk areas in Para state, the occurrence period and the illness distribution in space and time. Through a Knowledge Discovery in Databases approach a retrospective exploratory analysis was performed on SINAN LVH data between the years 2007 and 2019. The KDD process was applied as a data analysis methodology. The Global Moran's Indexes were positive along the period and higher in 2019 (0,45). The Local Moran’s I (LISA statistic) indicated that the municipalities with higher risk were in the northeast and southeast of state. The spatiotemporal scan statistics identified three circular clusters with a radius of up to 178 km, two of them occurred between 2007 and 2013 in the northeast, and the other one occurred between 2014 and 2019, in the southeast.