Search

Dairy farm hygiene audits were undertaken at 24 farms during summer and winter and the results compared with transformed bacterial indicator levels in raw milk samples collected during each audit. The bacterial indicators measured were total viable counts, Escherichia coli, coliforms, Bacillusspp., Bifidobacteriaspp., and Pseudomonasspp. The results of initial comparisons using Pearson product-moment correlation coefficients showed presumptive relationships between some bacterial groups and the subjective quantitative audit scores. When investigated further using linear regression, the presumptive relationships were found to be influenced by external factors. Possible reasons for the low correlations between on-farm hygiene and bacterial indicator counts in raw milk were further investigated. Measurements of the uncertainty associated with the bacteriological results were undertaken and revealed geometric relative standard deviations that ranged from 0.019 to 1.05. Toward the higher end of this scale, the uncertainty associated with the laboratory estimations of bacterial numbers may have been large enough to blur hygiene score-marker bacteria relationships. The samples obtained from on-farm raw milk storage tanks were representative of the whole tank contents and not a significant source of error. Although total bacterial counts are widely acknowledged by the milk industry as not always giving a true measure of on-farm hygiene during milking, we were unable to find any marker bacteria that showed consistently higher correlations and were thus better suited as indicators of on-farm hygiene.

Measurements of the conductivity of quarter milk samples were made in 31 cows in a 70‐cow herd in southeast England, for a period of 15 weeks. Over this period, 42 per cent of cow‐weeks and 20 per cent of quarter‐weeks had an increase in quarter milk conductivity of 10 per cent of more compared with the mean conductivity of the previous 14 milkings. Fourteen per cent of quarter‐weeks had an increase in conductivity of 15 per cent or more. The geometric mean somatic cell count (cell count) was higher in quarter‐weeks with a 10 per cent or greater increase in conductivity than in quarter‐weeks with a conductivity change of less than 10 per cent. At a conductivity threshold of 10 or 15 per cent and a cell count threshold of 200,000 or 400,000 cells/ml the specificity of this system was estimated to be 85 to 92 per cent, the sensitivity 40 to 54 per cent, the negative predictive value 87 to 93 per cent and the positive predictive value 33 to 55 per cent. The positive predictive value of the individual quarter milk conductivity was insufficiently accurate to be used as the sole criterion for the selection of quarters for early antibiotic treatment.