A comparison of the effects of different counting rules and aspect ratios on the level and reproducibility of asbestos fibre counts. Part 2: Effects on reproducibility. Final report on CEC Contract ‘Comparison of asbestos measurement methods at the workpl

The Commission of European Communities is currently preparing a Directive on airborne asbestos fibre regulations and required information on the effects of different counting criteria on the level and reproducibility of asbestos fibre counts to aid formulation of an associated reference method. This resulted in a three-stage investigation being initiated at a number of laboratories in different countries as part of a co-operative agreement between the Commission and the Canadian Government. An assessment of the effects on counting level was completed during 1982. This report describes the second stage of the programme which was designed to examine the effects on reproducibility of counts.Twnety-five microscopists from a total of sixteen laboratories in nine countries participated in the study. The counting rules used were those adopted by the Asbestos International Association (AIA) and a modified form of those developed in the United Kingdom Central Reference Scheme (CRS). Both counting criteria were considered in conjunction with the conventional minimum, fibre aspect ratio definition (viz. 3:1) and additionally the modified CRS rules were used in association with a minimum aspect ratio of 5 : 1. The samples evaluated were selected to embody a range of sample type. Details of the experimental design are presented in the report. The data is described and considered in terms of counting level, within-counter repeatability, within-laboratory variation and inter-laboratory variability.The AIA (> 3 : 1) package is the most widely known but was used routinely in only four of the participating laboratories. The modified CRS rules, being new, are not in use anywhere but 16 microscopists preferred them, 9 microscopists had no preference, and no microscopists expressed preference for the AIA rules.Differences in the counting levels associated with the counting tule packages varied from sample to sample and from laboratory to laboratory. Because of the subjective nature of the method, however, these differences are samll compared with the inter-laboratory differences observed when the same package is applied to the same sample. Several unexplained features emerged concerning the results obtained by a few laboratories.The modified CRS rules produced a marked improvement in reproducibility compared with the AIA rules. This was to be expected since the latter were developed somewhat arbitrarily, whereas the modified CRS rules were designed to minimise subjective-sources of inter-counter variability. The improved intra- and interlaboratory variation observed with both packages containing the modified CRS rules was particularly evident for asbestos cement and friction material samples. The smallest inter-laboratory differences occurred with the modified CRS rules used in conjunction with the 5 : 1 limiting aspect ratio; this package gave an overall coefficient of variation nearly half that associated with the AIA (> 3 : 1) package. However, even here, appreciable random and systematic variation in counts remained. In a practical sense, therefore, standardisation of counting criteria is much less important in support of European ‘control limits’ than the operation of a reference scheme to compare the performance of laboratories on a regular basis. “”