Techniques for analysis of disease clustering in space and in time in veterinary epidemiology

Abstract

Techniques to describe and investigate clustering of disease in space - the nearest-neighbour test, autocorrelation, Cuzick-and-Edwards' test and the spatial scan statistic - and in time - the Ederer-Myers-Mantel test and the temporal scan statistic - are reviewed. The application of these techniques in veterinary epidemiology is demonstrated by the analysis of a data set describing the occurrence of blowfly strike - both body strike and breech strike - between August 1998 and May 1999 in 33 commercial sheep flocks located within two local government areas of southeastern Queensland, Australia. By applying a combination of these methods, the occurrence of blowfly strike in the study area is well-characterised in both space and time. Guidelines for investigating disease clusters in veterinary epidemiology are discussed.