Statistical methods for geographical surveillance in veterinary epidemiology.

摘要: Spatial clustering and cluster detection are statistical analysis developed to address relevant scientific hypothesis. The difficulty stays in the large number of alternative hypothesis due different mechanisms that could generate anomalous cases aggregation. We review methods for marked point data (case/control) aimed describe spatial intensity disease risk, test randomness locate significant excesses. Bayesian Gaussian Exponential models used illustrate probabilistic aspects link with simpler non parametric tools shown. develop an informal guideline on faecal contamination dog parasitic diseases city Naples, Italy. Kernel density estimation resulted very sensitive bandwidth choice overemphasized localized excess, Ripley'K function Cuzick-Edwards were consistent each other while SatScan failed detect range was around 600 meters justifies several small clusters. powerful reconstructing phenomenon allow inference model parameters good agreement analysis.