Proficiency testing of water microbiology laboratories in The Netherlands.

In a 3-year period, four series of simulated water samples containing selected test strains were distributed to more than 50 laboratories in The Netherlands for bacteriological testing. Participating laboratories examined the samples by enrichment or membrane filtration methods, or both, for total coliform organisms, thermotolerant coliform organisms, faecal streptococci and standard plate counts (37 degrees and 22 degrees C) according to Dutch standard methods. The results were quantitatively satisfactory: the distribution of positive and negative results with subsamples conformed to stochastic variation; the standard deviation of membrane or plate counts was usually in the range which may be expected from a Poisson distribution, and there was good correspondence between average counts in participating laboratories and those expected from controls in the organizing laboratory. Problems of a qualitative nature were frequently encountered, however. Among them were a false positive response with a strain of Enterobacter cloacae in the thermotolerant coliform test; a false positive result with Clostridium perfringens in enrichment tests for total or thermotolerant coliform organisms and false positive results with Micrococcus varians in the faecal streptococcus test by membrane filtration. It is concluded that quality assessment should be a consistent activity in water microbiology laboratories. For this purpose, stable and well characterized reference materials are needed.