Statistical analysis of a micromethod in an interlaboratory test

A micromethod developed by Meier (1996) was assessed in an interlaboratory test with 7 labs supplying duplicate measurements on 5 to 9 days. This paper aims at evaluating the reproducibility and the repeatability of the micromethod. The effective concentration of microbicidal substances to reduce the number of test organisms by a factor of 10(5) was estimated from raw data of the microtiter test by means of a nonlinear regression model. Inclusion criteria were defined for the model fit and the results were checked for plausibility and for outliers according to methods prescribed by ISO 5725 (International Standard, 1994). From the remaining data (about 85% of the total data) inter-laboratory and intralaboratory reproducibility as well as repeatability were computed. The paper sets an example of how to analyze such interlaboratory tests. For practical purposes a table shows confidence limits of the relative precision for a broad range of interlaboratory test designs. In addition, test designs are given which lead to statistically qualified decisions about acceptance or rejection of a claimed effective concentration of a microbicidal substance. The results clearly qualify the micromethod for practical application in testing microbicidal substances.