Assessing the effect of Aedes (Stegomyia) aegypti (Linnaeus, 1762) control based on machine learning for predicting the spatiotemporal distribution of eggs in ovitraps

Abstract

Background: Aedes aegypti is the dominant vector of several arboviruses that threaten urban populations in tropical and subtropical countries. Because of the climate changes and the spread of the disease worldwide, the population at risk of acquiring the disease is increasing. Methods: This study investigated the impact of the larval habitats control (CC), nebulization (NEB), and both methods (CC + NEB) using the distribution of Ae. aegypti eggs collected in urban area of Santa Bárbara d'Oeste, São Paulo State, Brazil. A total of 142,469 eggs were collected from 2014 to 2017. To verify the effects of control interventions, a spatial trend, and a predictive machine learning modeling analytical approaches were adopted. Results: The spatial analysis revealed sites with the highest probability of Ae. aegypti occurrence and the machine learning generated an asymmetric histogram for predicting the presence of the mosquito. Results of analyses showed that CC, NEB, and CC + NEB control methods had a negative impact on the number of eggs collected in ovitraps, with effects on the distribution of eggs in the three weeks following the treatments, according to the predictive machine learning modeling. Conclusions: The vector control interventions are essential to decrease both occurrence of the mosquito vectors and urban arboviruses. The inference processes proposed in this study revealed the relative causal impact of distinct mosquito control interventions. The spatio-temporal and the machine learning analysis are relevant and Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation robust analytical approach to be employed in surveillance and monitoring the results of public health programs focused on combating urban arboviruses.