Several vector-borne illnesses, including West Nile virus, Lyme Disease, dengue, and chikungunya have seen dramatic geographical spread in recent decades, sparking a need to alert residents with timely and empirically informed warnings on the current and future risk of exposure to such illnesses. The Prediction System for Vector-Borne Diseases (PresVBD) combines the flexibility of modern computing and the intelligence of a location specific science-based understanding of vector-borne diseases (VBD) to provide early warning for improved public health response. Mosquitoes, ticks, and other disease vectors are biologically linked to temperature, rainfall, vegetation, and other features of the natural and built environment, so PresVBD focuses especially on tools that enable on-demand data retrieval for dynamic, spatially refined environmental data. Our approach to estimate VBD risk is to use long-term time series health outcome data and generalized linear regression to develop location specific weekly predictions of the potential for West Nile virus or other VBDs. Temperature and rainfall conditions are essential indicators, also play important roles. These models are quite successful in data-rich regions, but are weaker in places where data are sparse. We have developed the PresVBD to enable recalibration of a model's coefficients based on a user initiated selection of data by location, for more limber development of VBD prediction in local regions. PresVBD is a modular system with tools designed to grab data from several sources, process the data to create weekly, location-specific indicator variables, and then run a new model or make a prediction from an existing model. The system uses integrated technical resources: Clowder, BrownDog, GeoDashboard, and DataWolf, developed at the National Center for Supercomputer Applications. These are employed to allow a research team to provide better information to public health and vector control decision-makers on the risk of VBDs.
|Original language||English (US)|
|Title of host publication||AGU Fall Meeting Abstracts|
|State||Published - 2018|