Laboratory induced variations in a standardized Salmonella isolation method.

Four microbiologists from four European laboratories came together to a single laboratory for two weeks and during that period examined 100 samples of minced meat artificially contaminated with salmonellae and 100 naturally contaminated samples using a standardized technique. No significant differences were found between the results obtained by the different analysts with 92 to 97 of the artificially contaminated and 30 to 42 of the naturally contaminated samples being found positive. Over a further period of eight weeks, these analysts examined the same number of samples in their own laboratories again using an identical procedure. There were no significant containing approximately 100 salmonellae per gram but significant differences (43 to 93) were found in isolation rates from the artificially contaminated samples copntaining approximately 1 salmonella per gram. The differences in findings between the first and the second part of the experiments may have been caused fairly by variations in media preparation, temperature of incubators, storage conditions of media and thawing conditions of samples etc., and secondly by differences in working conditions such as performing only one task during two weeks in contrast to examining the samples as part of the daily routine. At high levels of contamination, these factors do not appear to have an important influence on the performance of a laboratory.