Page:NIOSH Manual of Analytical Methods - Chapter L.pdf/6

 Crawford investigated the effect of some of these factors and estimated most of them made a difference in the final count on the order of 20% [25]. Many of the methods currently in use have slight variations in their interpretation of which fibers to count and thus can contribute to variation in results between countries and organizations. NIOSH Method 7400 contains two sets of counting rules, the A and the B rules. The A rules are used for asbestos and are consistent with counting rules in previous NIOSH methods. The A rules are required for asbestos counting by OSHA because of legal precedent in regard to the 3:1 aspect ratio rule. The A rules do not have an upper diameter limit for fibers to be counted. The B rules were introduced as an alternative to the A rules when Cowie and Crawford found that these rules agreed best with previous PCM counts, yet had improved precision [25]. The B rules have been informally adopted for use with fibers other than asbestos because these rules include the upper diameter limit of 3 :m. This upper diameter limit significantly reduces the counting of typically large-diameter fibers, e.g., glass and cellulose, that are unlikely to deposit in the lungs [26]. c. Quality assurance schemes Sample recounts. Most methods require individual counters to recount about 10% of the field samples to ensure consistent counting procedures and alert the analyst in the case of problem samples. It is also recommended that counters have samples that are routinely recounted to ensure consistent counting within a laboratory over time. One of the difficulties in analyzing errors made by analysts during PCM counting is that individual fields are difficult to relocate after the analyst has finished counting a slide. Differences in counts between analysts have often been ascribed to local variations in loading on the filter. Pang’s development of a slide coverslip that defines counting areas on the sample solves this problem [17]. Areas on the coverslip are vacuum coated with a thin layer of gold and platinum using an electron microscope grid as a mask. This leaves defined areas on the coverslip that can be located by grid index marks. Thus, specific fields in a sample can be readily located. Using this grid mapping approach, the location, orientation and shape of each fiber can be noted and differences in counts can be reconciled on a fiber-by-fiber basis. The coverslips have been used to study fiber counting accuracy by comparing routine counting of specified fields to counts agreed upon by a group of competent counters. It was found that the principal errors for chrysotile fiber samples were due to missing fibers close to the visibility limit, while the principal errors for amosite fiber samples were caused by incorrectly sizing fiber length near the 5-:m limit. The chrysotile samples were therefore typically undercounted (negative bias), while the amosite samples had increased variability with individual counters being biased either high or low. Both these errors can be reduced by training counters with pre-counted reference slides prepared using Pang’s coverslips [17] (Omega Specialty Instrument Co. Chelmsford MA). In addition, these reference slides can be used on a routine basis to ensure consistency in counting. These coverslips or modified versions show great promise for training analysts and perhaps for improving quality assurance schemes. Interlaboratory sample exchanges. Crawford et al. found that use of sample exchange programs was more important in ensuring agreement between laboratories than similarity in details of the counting rules [27]. Thus, exchange of field samples between laboratories is

3/15/03

148

NIOSH Manual of Analytical Methods