Biostatistics Seminar: Chris Barker, Microbiology & Immunology, UC Davis
DATE: Tuesday, April 30th, 2013
LOCATION: 1047 Gladys Valley Hall (map)
SPEAKER:Chris Barker, Dept. Pathology, Microbiology, & Immunology, School of Veterinary Medicine, UC Davis
TITLE: Combining data and models to enhance surveillance for West Nile virus
ABSTRACT: West Nile virus (WNV) is the most important and widespread mosquitoborne pathogen affecting human health in the U.S., having caused > 36,000 diagnosed cases of human disease since its arrival in New York in 1999, including > 5,000 during a large outbreak in 2012. Transmission of WNV is complex, highly seasonal, and involves multiple mosquito and bird species that amplify the virus in natural cycles, which can result in spillover transmission when infectious mosquitoes bite humans.
In California, WNV-related disease risk is monitored by a cooperative surveillance network involving mosquito control districts, public health agencies, and UC Davis. This system has proven effective for estimating concurrent risk and guiding outbreak response, but earlier warning – weeks- to seasons-in-advance – is needed for proactive interventions to avoid epidemics. This presentation will describe development of hierarchical models to address several challenges for estimation and prediction, namely multicollinearity in seasonally varying environmental covariates, spatial and temporal dependence, zero inflation, and posterior predictive checking for model validation. Goals for further development will be discussed.