| Abstract: | When applying methods for the detection of pathogenic micro-organisms in foodstuffs, information on the distribution of the target organism in the foodstuff submitted to a test should be available. The sampling plan used should allow to detect the presence of low levels of the target organism with a high probability. The individual steps of a detection procedure (pre-enrichment, selective enrichment, isolation and confirmation) need careful examination. It is important that inoculation of low levels of the target organism leads to successful enrichment even in the presence of relatively large numbers of competing organisms. In cases where competing micro-organisms form suspect colonies, there is a risk that false-negative results are obtained, because colonies of the target organism may not be isolated. Collaborative trials have to be carried out to assess the performance of presence/absence tests. Meaningful results are obtained only, if the test samples contain low levels of the target organism and if the effect of competing micro-organisms is checked. While it can hardly be disputed that the determination of the sensitivity and specificity of the test method provides valuable information, this cannot be said for "accordance" and "concordance", two recently proposed parameters which correspond to repeatability and reproducibility in quantitative tests. A better alternative may be to specify the probability to obtain two positive results, when analysing samples containing the target organism under repeatability or reproducibility conditions. In an analogous way, the probability to obtain two negative results with samples containing competing micro-organisms, but not the target organism, could be specified. |
