Eco-epidemiology of West Nile virus in the midwestern United States

Infectious diseases are on the rise globally. Although only accounting for 17% of all infectious diseases, vector-borne diseases are increasing the fastest. Global changes in climate, particularly in precipitation and temperature, directly affect the life cycle of arthropod vectors, including the incubation period for any pathogen they transmit. Mosquitoes are among the most responsive to climatic changes, and several species are considered the most successful invasive organisms on our planet. In 1999, West Nile virus (WNV) arrived in New York City and caused a local outbreak affecting birds and humans. In four years, the virus reached California and was present in nearly every state. Two decades later, WNV is now the most important mosquito-borne disease in North America and continues to cause human infection and death. Countless resources, person-hours, and millions of dollars have been used to control and monitor WNV throughout the country. However, predicting when and where infection will occur has proven immensely difficult, largely due to the complex relationships of numerous interacting factors affecting WNV disease ecology. Considerable variation in factors affecting WNV transmission derive from climatic, environmental, physical, and human socio-economic and demographic forces, and can vary weekly and across fine-scales. Several research teams have investigated these dynamics, attempting to improve our understanding of the drivers of WNV. While research is improving many aspects of mosquito control and mitigating risk, human infection continues. Additionally, results and interpretations of WNV are less clear, sometimes conflicting among overlapping study areas. The state of Illinois, experiencing the first positive mosquitoes in 2001, has since produced the fifth most human cases in the country. Fortunately, the Smith and O’Hara Ruiz labs at the University of Illinois have dedicated over 10 years of research in improving our understanding of transmission in the Chicago region. Several key findings generated from these labs include the link between housing age and infection, key associations derived from complex spatial epidemiology methods, and human activity and behavior risk. The O’Hara Ruiz lab was also among the first research groups to associate strong relationships between mosquito infection with temporal lags in precipitation and temperature. Largely expanding from previous efforts in our lab, the aims of this dissertation are to create robust and accurate WNV forecast models for the midwestern United States. Specifically, the chapters of this dissertation focus on very fine-scale drivers of disease, and then comparatively analyze best-fit models across scales in Chicago. Lastly, this dissertation expands efforts from Chicago to the Midwest, evaluating mosquito infection across 118 counties and 8 states. Findings from this research demonstrated that drivers of WNV vary by scale. Specifically, the finer the scale, the more important an included covariate becomes. However, as scales become broader, the overall performance of WNV models increases. This research also found that precipitation and temperature (and their respective 2 and 3 week lags) are consistently the most important covariates in WNV transmission, corroborating several studies. However, acquiring additional data sources does improve model strength, but the overall net benefit given allocation of resources may not be the most efficient use of time and effort. Lastly, this study found that the upper Midwest, overall, has adequate spatial coverage of mosquito infection through current surveillance practices. However, there are several notable gaps in our understanding of disease, particularly in most of Iowa and southern Wisconsin. Additionally, mosquito infection is increasing by about 14.2% annually, and is not an artifact of increased mosquito control efforts. Notable human outbreaks in 2005-2006, 2012, and 2018 coincide with years of highest mosquito infection in the Midwest. The time between these outbreaks is decreasing by about 1 year. The knowledge gained from this dissertation provide important, but sobering insights into the projections of human WNV infection in the Midwest. There is little doubt that future outbreaks will not only occur, but increase in frequency and numbers of infected. The models created and compared in these analyses can be used to understand trends and forecasts of WNV in other regions of the United States, as well as with other mosquito-borne diseases.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

VMS - Pathobiology

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2021

Author dc:creator

• Uelmen Jr, Johnny Albert

Contributors dc:contributor

• Smith, Rebecca
• Jarosinski, Keith
• Allan, Brian
• Fraterrigo, Jennifer
• Ryan, Sadie
• Patz, Jonathan

• West Nile virus
• eco-epidemiology
• disease modeling
• Culex, climate change

Statement dc:rights

• Copyright 2020 Johnny Uelmen

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/109335

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/109335