Estimates of the prevalence of transmissible spongiform encephalopathies in sheep and goats in France in 2002.

Abstract

An active surveillance programme for transmissible spongiform encephalopathies (TSES) in sheep and goats was implemented in France in 2002 at abattoirs and rendering plants. The analysis of the results of this programme highlighted three biases: a potentially non-random sampling scheme in both rendering plants and abattoirs, a heterogeneous geographical sampling ratio, and the use of two diagnostic tests of unequal sensitivity. Simulations were run to estimate the prevalence of TSES by taking these biases into account. A comparison of the prevalence of TSES calculated from the raw data with the simulation results showed that the effects of non-random sampling were minor in comparison with the effects of the heterogeneous geographical sampling ratio and the use of two diagnostic tests.