The Epidemic Next Door: A statistical framework for understanding current and future arbovirus outbreak risk

Abstract

Arboviruses, or viral infections transmitted from arthropods to humans, play an outsized role in human outbreaks, as many recent zoonotic outbreaks (e.g., West Nile virus from 1999-2014, chikungunya virus from 2004-14, and Zika virus from 2007-17) have arboviral origins. Furthermore, the widespread nature and threat of arboviruses such as dengue, for which roughly half of the global population lives in an area environmentally suitable for transmission, calls for increased attention as to why arbovirus outbreaks have remained pervasive despite rapid infrastructure, vector control, and global surveillance improvements over the past 50 years. This thesis aims to understand the climatic and anthropogenic factors most salient in arbovirus outbreaks. Using dengue as a case study, I advance a novel technique for mapping outbreak risk using machine-learning based species distribution modeling (SDM). By comparing current dengue risk to the distribution of its main vector, the Aedes aegypti mosquito, I find that dengue outbreaks follow a similar spatial pattern as its vector, with differences in the importance of population density underscoring the role of anthropogenic factors in amplifying zoonotic spillover events into outbreaks. The results of this study suggest that while climate change and continued globalization may increase the risk of arbovirus outbreaks globally, SDM techniques present a viable method to understand the threat of these outbreaks and can therefore be used to develop more effective and efficient vector control and disease eradication programs and policies.