Setting control limits for Escherichia coli counts in samples collected routinely from pig or beef carcasses

Records of Escherichia coli counts in samples routinely collected from carcasses were obtained from one pork and three beef packing plants. The data obtained from each plant were divided into sets from consecutive 6-month periods. For each set of counts, log total counts were calculated for subsets of various sizes. For each set of log total counts, the mean (x), the standard deviation (SD), and an action limit of x + 3 SD were calculated, and the set was tested for a normal distribution. With the data from samples collected at the pork packing plant during 6 years, the proportion of samples with counts of zero in the 12 sets ranged from 15 to 45%. For that plant, appropriate action limits could be derived from log total counts for subsets of nine unit values. With the data from samples collected during 8 years at one beef packing plant, the proportion of samples with counts of zero in the 16 sets ranged from 88 to 99%. For that plant, appropriate action limits could be derived from log total counts for subsets of 15 unit values. With the data from samples collected during 2 or 2.5 years at each of the other beef packing plants, the proportion of samples with counts of zero in all sets was > 99%. For those data, action limits could not be derived from values for subsets of log total counts.