Factors associated with variation in bulk-tank-milk Salmonella Dublin ELISA ODC% in dairy herds

Our objective was to determine factors that contribute to variation in bulk-tank-milk Salmonella Dublin enzyme-linked immunosorbent assays (ELISA) corrected optic-density measurements (ODC%) in dairy herds. We constructed hierarchical mixed models with repeated bulk-tank-milk ELISA ODC% in 31 Danish dairy herds. Four models included different combinations of explanatory factors, and we compared how well these models described the variation in the data. Herd was included as a random effect nested within Salmonella status and barn type. Detection of Salmonella Dublin or Salmonella Typhimurium by bacteriological culture of individual faecal samples or of slurry samples was associated with higher bulk-tank-milk ELISA ODC%, as was apparent Salmonella prevalence, the mean ELISA ODC% or mean-yield-corrected ELISA ODC% in milk samples collected from all individual cows. However, combinations of risk factors that included number or prevalence of cows with a very high ELISA ODC% provided better models, indicating that the effect of the cow-level explanatory variables on the bulk-tank-milk ELISA ODC% was related to the activity of the infection in the herd. Barn type (loose housing or tie stalls) was not associated with the variation in bulk-tank-milk ELISA ODC% in these models, which might be useful in planning of surveillance programs and intervention strategies. (c) 2005 Elsevier B.V. All rights reserved.