Application of one-list capture–recapture models to scrapie surveillance data in Great Britain

Abstract

In this paper, we apply one-list capture–recapture models to estimate the number of scrapie-affected holdings in Great Britain. We applied this technique to the Compulsory Scrapie Flocks Scheme dataset where cases from all the surveillance sources monitoring the presence of scrapie in Great Britain, the abattoir survey, the fallen stock survey and the statutory reporting of clinical cases, are gathered. Consequently, the estimates of prevalence obtained from this scheme should be comprehensive and cover all the different presentations of the disease captured individually by the surveillance sources. Two estimators were applied under the one-list approach: the Zelterman estimator and Chao's lower bound estimator. Our results could only inform with confidence the scrapie-affected holding population with clinical disease; this moved around the figure of 350 holdings in Great Britain for the period under study, April 2005–April 2006. Our models allowed the stratification by surveillance source and the input of covariate information, holding size and country of origin. None of the covariates appear to inform the model significantly.