Climate and environmental drivers of spatio-temporal spread of arboviruses
Conference
64th ISI World Statistics Congress - Ottawa, Canada
Format: IPS Abstract
Keywords: "bayesian, climate, hierarchical, socioeconomic
Session: IPS 194 - Frontiers in Data Science, Health, and the Environment
Thursday 20 July 2 p.m. - 3:40 p.m. (Canada/Eastern)
Abstract
Dengue and Zika viral infections are Aedes mosquito-borne diseases, caused by viruses which are members of the Flaviviridae family. Despite a substantial body of research, a few studies have attempted to probabilistically model the two diseases simultaneously. In this study, we present their joint analysis across Brazil, from January 2015 to June 2019, taking advantage of multiple data sources, including satellite-derived information on the local vegetation, re-analysis data on climate, census-based data, and population characteristics. Then, by relying on a hierarchical approach grounded in a Bayesian framework, we build a multi-likelihood model for monthly disease counts, using as spatial unit of analysis the 5570 Brazilian municipalities, and we quantify the risks associated with climate, local environmental and socioeconomic conditions, while accounting for latent spatio-temporal patterns. Preliminary analyses show that the risk of disease is unequally distributed, both geographically and socioeconomically and present different temporal patterns across the Brazilian regions.