An evaluation of screening programs for the detection of brucellosis in dairy herds.

Data were collected from approximately 1000 dairy herds, initially blood tested for brucellosis in 1977, in each of southeastern and southwestern Ontario. These data were used to evaluate three brucellosis screening programs, the area recertification program, the market cow program and the milk test program. The milk test program was the most efficient program at detecting brucellosis, 29.5% of the herds tested were classified as infected, but lacked the ability to detect a large proportion of "infected" herds (relative sensitivity = 24%). The market cow program was more efficient than the area recertification program at finding infected herds, 3.9% of the herds tested under the market cow program were infected, but had a low relative sensitivity of 12%. The area recertification program was least efficient, 2.3% of herds tested under the area recertification program were infected, but had the highest relative sensitivity (53%). The relative efficiency (predictive value) of the programs was not affected significantly by location of the herds, season of the initial test or herd size. The relative sensitivity of the milk test program was significantly higher in eastern than western Ontario and tended to decrease as herd size increased. The market cow program tended to be more sensitive in the summer months. The relative specificity of the milk test program (0.997) was higher than that of the market cow program (0.960) and the area recertification program (0.884).