and the type and age of the animals. Secondly, after accounting for the bovine TB status of the slaughtered animals, other factors likely to affect the efficiency of surveillance include physical factors, such as line speed and light intensity, and human factors, such as the quality of inspection, as influenced by the competence of the inspector. Several preliminary studies have been made, some based on univariable analyses (de Kantor and others 1987, Corner and others 1990, Kobe and others 2000, Lenehan and others 2000 and one on a multivariable analysis (Martin and others 2003). The multivariable approach is preferable because it ensures that measures of efficiency are adjusted for the likely bovine TB status of the animals slaughtered at each factory. Martin and others (2003) detected a seven-fold difference between factories in the rate of disclosure of lesions among attested cattle at slaughter, after controlling for a limited range of potential confounding factors, including the month and year of slaughter, and the class of cattle.The aim of this study was to determine the relative efficiency of factories in disclosing lesions in attested cattle, on the basis of the submission of lesions and their confirmation as bovine TB. The analyses have attempted to control for a broad range of potential confounding factors, including the geographical risk of infection, the bovine TB history of the farms of origin, the animals' characteristics (age, sex and whether they were homebred) and the season of slaughter. The study is based on data relating to lesions submitted and confirmed as bovine TB during 2003 and 2004.
MATERIALS AND METHODS
DatabasesThe data were obtained from the Centre for Veterinary Epidemiology and Risk Analysis at University College Dublin, Ireland. Three primary databases were used, as follows.
BOVINE tuberculosis (TB), caused by the bacteriumMycobacterium bovis, is a chronic infectious disease affecting cattle, other animal species and human beings. A national eradication programme was initiated in Ireland in 1954 and initially made considerable progress, but further advancement has proved difficult. The programme consists of annual tuberculin skin testing of all cattle; reactors are removed and the herd is restricted from trade until all the animals test negative on two consecutive occasions two months apart (Good and others 2003). The detection at slaughter of gross lesions in attested cattle, that is, cattle from herds considered disease free on the basis of annual skin testing (factory surveillance), is an additional and important method for detecting infected herds as part of the programme. Between 1993 and 2001, between 27 and 46 per cent of all new herd breakdowns in any year were detected by factory surveillance (O'Keeffe and White 1999; T. Clegg, personal communication). Routine factory surveillance, conducted in Ireland by veterinarians employed as temporary veterinary inspectors, is based on the palpation, incision and inspection of a defined range of lymph nodes, including the parot...