Evaluation of errors and limits of the 63-microuse-dust-fraction method, a surrogate to predict hidden moisture...

The aim of this study is to analyze possible random and systematic measurement errors and to detect methodological limits of the previously established method.

To examine the distribution of random errors (repeatability standard deviation) of the detection procedure, collective samples were taken from two uncontaminated rooms using a sampling vacuum cleaner, and 10 sub-samples each were examined with 3 parallel cultivation plates (DG18). In this two collective samples of new dust, the total counts of Aspergillus spp. varied moderately by 25 and 29% (both 9 cfu per plate). At an average of 28 cfu/plate, the total number varied only by 13%.For the evaluation of the influence of old dust, old and fresh dust samples were examined. In both cases with old dust, the old dust influenced the results indicating false positive results, where hidden moist was indicated but was not present. To quantify the influence of sand and sieving, 13 sites were sampled in parallel using the 63-mum- and total dust collection approaches. Sieving to 63-mum resulted in a more then 10-fold enrichment, due to the different quantity of inert sand in each total dust sample.

The major errors during the quantitative evaluation from house dust samples for mould fungi as reference values for assessment resulted from missing filtration, contamination with old dust and the massive influence of soil. If the assessment is guided by indicator genera, the percentage standard deviation lies in a moderate range.