Use of a Bayesian risk-mapping technique to estimate spatial risks for mare reproductive loss syndrome in Kentucky

Abstract

To estimate spatial risks associated with mare reproductive loss syndrome (MRLS) during 2001 among horses in a specific study population and partition the herd effects into those attributable to herd location and those that were spatially random and likely attributable to herd management. Animals-Pregnant broodmares from 62 farms in 7 counties in central Kentucky. Veterinarians provided the 2001 abortion incidence proportions for each farm included in the study. Farms were georeferenced and data were analyzed by use of a fully Bayesian risk-mapping technique. Large farm-to-farm variation in MRLS incidence proportions was identified. The farm-to-farm variation was largely attributed to spatial location rather than to spatially random herd effects. Results indicate that there are considerable data to support an ecologic cause and potential ecologic risk factors for MRLS. Veterinary practitioners with more detailed knowledge of the ecology in the 7 counties in Kentucky that were investigated may provide additional data that would assist in the deduction of the causal factor of MRLS via informal geographic information systems analyses and suggest factors for inclusion in further investigations.