Document p06n37ERwG1xb7M7Zz2XMw4E

22690 Federal Register / Vol. 51, No, 119 / Friday, June 20, 1986 / Rules and Regulations the ALA'S CV estimate for intersample method reflects a truer estimate of the asbestos fibers on a filter, the practical . variability necessarily reflectsa random random variability of the method limit of detection for the NIOSH 7400 error that is intrinsic to the P&CAM 239 because, of the greater degree of method is much higher: or NIOSH 7400 methods, finding instead that nonrandom error, caused by inadequate sampling times, low fiber counts, etc. accounts for what the A1A consistency in laboratory practices and training between the different johnsManville laboratories that generated the sampling data. NIOSH's estimate of the . . . The practical limit of reliable detection incorporates important practical factors that prevent the theoretical detection limit from being achieved such as reasonable terms intersample variability. OSHA overall precision of the phase contrast sampling times, reasonable pumping rales thus rejects the argument that method of asbestos analysis are similar and filter loading with nuisance dust. intersample variability may be a more to the estimates reported by Ogden (Ex. Problems with nuisance dust loading will significant source of sampling and 84-448), who also studied analytical vary from one atmosphere to another and analytical error than intra- or interlaboratory errors. Summary of Evidence on Sampling and Analytical Variability results from laboratories that implement similar training and laboratory practices. OSHA believes that, by promulgating requirements for specific - monitoring procedures and quality may seriously limit the range of the method. . . . A number of measurement methods in the record have suggested that practical limits of reliable detection fall in the range or 0.1 f/cc to 0.5 f/cc. In fact. P&CAM 239 establishes 0.1 In the preceding discussion. OSHA control measures, it is possible to limit f/cc. as its lower bound even though reviewed the rulemaking evidence sources of sampling and analytical error exposure levels well below this may be describing the sources and extent of sampling and analytical error associated with measuring airborne asbestos' concentrations using the NIOSH P&CAM 239 or NIOSH 7400 methods. The major sources of sampling and analytical variability described by-the AIA (Ex. 328} include intralaboratory variability, interlaboratory variability, to those that are random. Therefore, OSHA finds that NIOSH's estimates for the overall precision of the P&CAM 239 method, and consequently for .the NIOSH 7400 and OSHA Reference methods, are the most reliable estimates in the record of the random sampling and analytical variability of these methods. As.long as samples are obtained using the method. NIOSH's new Method 7400 does claim a lower value. 0.02 f/cc. but made clear that this is only a theoretical limit of detection by stating that it applies only in the absence of excessive nuisance dust loading. NIOSH has not explained why it chose to depart in Method 7400 from the practical limit of reliable detection employed in-P&CAM 239. Considering the total absence in the record and intersample variability. Together, obtained with a fiber density of 80-100 of any published testing results on Method AIA estimated that these sources of fibers per 100 fields counted. OSHA 7400. there has never been any variability result in an overall CV for the believes that the standard analytical . demonstration that it is practical to achieve NIOSH P&CAM method ranging from 0.45 to.0.65 (Ex.. 328, p. A-23). Based on this estimate for total CV, the AIA error (SAE) of the ORM will be between 18.9 and 21.3 percent, with an upper estimate of 24.7 percent, based on the the claimed value in manufacturing workplaces (Ex. 128, pp. A-24 through A-27). argued that OSHA must recognize that this unique variability limits the degree to which the asbestos PEL can be reduced, calls for more than usual enforcement flexibility, and. at the same time, assures that workplaces in compliance with a PEL will keep average exposures much lower than the PEL.'(Ex. 328, p.A-1) The AIA further argued that, given the high variability of the NIOSH P&CAM 239 and NIOSH 7400 methods, analysis of Dr. Ogden. This SAE coincides With the SAE of 25 percent currently used by OSHA for evaluating compliance samples of.airborne asbestos. Therefore. OSHA. will retain the SAE for asbestos at 25 percent. The Agency also finds that use of the ORM will not require employers to reduce their employees' exposures to levels substantially below the PEL of 0.2 f/cc to ensure that they are in compliance with the new PEL. Dr. Chatfield also expressed this view al the hearing, stating that the higher flow, rates permitted by the NIOSH 7400 method ". . . will result in (the] capture of even larger pieces of nuisance dust than are collected currently using P&CAM 239 adding further to the filter obstruction problem" (Tr7/6, p. 64). In his testimony, Dr. Taylor staled that NIOSH intended the higher flow rates permitted by the 7400 method to be used for taking clearance samples and ' promulgation of a 0.2 f/cc PEL would not leave an adequate margin between the PEL and the "practical limit of detection" (Ex. 328, p. 11-18}, and that employers would thus not be able to reliably determine whether'their employees' exposures are in excess of the PEL. OSHA rejects these arguments for two reasons. First, OSHA does not agree with the AIA's estimate of the overall precision of the NIOSH P&CAM 239 . method. As discussed in the section above, the data relied upon by the AIA do not predominantly reflect random sources of sampling and analytical error instead, they reflect nonrandom error caused by the use of different monitoring methods and quality control programs by laboratories participating in the study. On the other hand, the study: by Busch et al. (Ex. 84-62} of the precision of the NIOSH P&CAM 239 Reliable Limit of Detection not for the routine monitoring of airborne asbestos levels in workplaces NIOSH has reported that the 7400 (Tr. 6/21, p. 181). OSHA agrees that the -method for asbestos sampling and higher flow rates permitted by the 7400 analysis has a reliable limit of detection method may contribute to filter overload ' of 0.02 f/cc, based on collecting a 1,920- or interference by other particles, and liter sample (i..e,, collecting an 8-hour has therefore limited the ORM's flow sample using a flow rate of 41pm) and ' rate to 2.5 lpm. Although OSHA believes obtaining the minimum acceptable fiber that limiting flow rate will, in the great density of 80 fibers per.100 fields. Using majority of cases, eliminate the filter the formujas for calculating limits or detection (Exs. 84-444, 84-478), OSHA overload problem, the ORM does permit the use of a 37-mm diameter filter in has determined that the reliable limit of specific instances where fitter detection for the ORM is 0.03 f/cc/, overloading may be a problem. based on obtaining a fiber density of 80 fibers/100 fields from a 1,200-liter sample (2.5 lpm over 8 hours); the limit of detection for the ORM is thus well below the b.l f/cc action level included in the revised standards for general industry and construction. . . However, since the use of the 37-mm diameter filler doubles the limit of detection that can be achieved with the smaller filter, the larger filter may only be used if employers provide a written . justification for its use. OSHA believes, based on testimony presented by Drs. The AIA argued that, because of the. Ogden and Taylor to the effect that problem of nuisance dust obscuring nuisance dust overload is a rare T'.T; i' -.VlTf' GLEASON-000938