Pets, People and Weather: Forecasting the Emergence of Infectious Disease in Your Broadcast Area

Vector-borne diseases agents require an insect host for propagation and/or dissemination and are responsible for approximately 20% of the global burden of infectious disease. These diseases are often transmitted when blood-feeding arthropods, such as flies, fleas, or ticks, deliver infectious pathogens to mammalian hosts while feeding. Common vector-borne diseases, many of which are zoonotic, include Lyme disease, anaplasmosis, ehrlichiosis, canine heartworm, West Nile virus, leishmaniasis, plague, and malaria. The World Health Organization estimates that over half of the world's population is currently at risk of infection by vector-borne disease. Substantial challenges exist in our ability to prevent and/or control outbreaks of vector-borne disease. Insects carrying disease can move considerable distances and be introduced to new geographic areas by human travel; war and famine; international commerce; animal movement; migratory birds and encroaching urbanization. Beyond human and animal-related factors, we now know rainfall, heat, humidity and wind patterns are key determinants for vector-borne disease spread. The complex epidemiology of vector-borne disease makes surveillance, early detection and prediction of emergence the most viable solutions to control vector-borne disease. To address the growing need for surveillance and early detection, the Companion Animal Parasite Council (CAPC) has developed a mathematical model to forecast changes in canine disease prevalence with some of the diseases examined, e.g., the tick-borne diseases, being important disease agents of both humans and dogs in the US. Our methods are based on a massive data set, compiled since 2011, that contains millions of annual test results for diseases from veterinary practices across the US. The statistical methodology uses Bayesian spatio-temporal models to relate putative factors (e.g., climate conditions, socioeconomic characteristics, local topography, and vector distribution) to county-level diagnostic test results. The maps have focused to date on three bacterial diseases transmitted by ticks, i.e., Lyme disease, anaplasmosis, ehrlichiosis, and the canine heartworm, caused by the nematode Dirofilaria immitis transmitted by mosquitoes. The tick-transmitted diseases are the same that occur in people in the same geographic areas of the United States. Canine heartworm does occasionally infect and cause disease in people, but the diagnosis of such infections is rare. Earlier collections of a smaller set of county-by-county canine Lyme disease prevalence data collected from the same sources as that on the CAPC maps has been shown to be suggestive of human risk for Lyme in an area. That is, when more than 5% of dogs in a given area are reported as positive for Lyme disease, cases will begin to appear in people in the same county. The CAPC goal remains to forecast the risk of a dog presenting to a veterinarian's office with one of these and other infectious diseases, however, it is likely that others will continue to focus on the relationship between infections in pets and what they can tell us about associated infections in people. During the presentation, we will illustrate the robustness of our approach through comparison of forecasted vs. actual disease prevalence rates for the year 2015. Importantly, forecast maps that are easily understood by both medical/veterinary practitioners and the general public will be presented. In the end, the goal is to present a series of maps for major metropolitan areas that will present the risk different risk each month of a dog testing positive for one of the several different diseases for which the data is being collected.