Document Z9ddOqv0B7MkJr9w1DDBobZJ

Federal Register / Vol. 51, No. 119 / Friday, June 20, 1986 / Rules and Regulations 22637 significant impact of local lung cancer death, rates, assuming an overall lung cancer SMR of 200 or more for these workers. The effects of shipyard and asbestos plant employment make the use of local death rates inappropriate for this study [Ex. 84-35, p. 879880]. In addition, state (South Carolina] mortality rates from lung cancer were similar to those of the United States. Moreover. "(AJvailable smoking data for this cohort suggest that the observed lung cancer and nonmalignant mortality excess among white males cannot be explained by cigarette smoking independent of asbestos exposure" (Ex. 84-37. p. 430). Although Crump pointed to the arguments raised by Acheson and Gardner [Ex. 84-243] lhat local rates should have been preferred. OSHA found these arguments unconvincing. ' Crump recommended a KL of 0.023. approximately half the value of KL calculated by OSHA. Crump noted that: * * * Not only does this modification provide a better Hi to the Dement et al. data, the estimated background rate agrees closely with the 75% excess of local lung cancer rates over national rates (See Figure 3 of Acheson and Cardner. 1983 [Ex. 84-243]). The lower estimate of KL = 0.023 also reduces the discrepancy between this and other studies which show a much smaller KL. OSHA believes that a reduction of the Kl to 0.023 is inconsistent with the available data: First, Dement et al. noted that: "* even if rates for contiguous counties - had been used ... the expected lung cancer rates for white males would have been increased by only, approx. 15%. not nearly . sufficient for the observed excess lung cancer risk (Ex. 84-35. p. 880], Moreover, as Dement pointed out in 1982: * * 1 rates for contiguous counties for black males were approximately 45 percent below U.S. rates: thus, the overall excess, among blacks is underestimated by the present study, although the numbers were small (Ex. 84-229. p. 179). Thus, to some extent, these overall estimates may beunderestimated. Hence, OSHA concludes that its original estimate of KL for this study, 0.042, is valid arid reasonable, and thus has . adopted it for the final rule. C.jCalculation of the Overall KL. OSHA's best estimates of KL from the proposed rule, and the final determination of KL for each study are given in Table 2, along with a range of uncertainty. The, ranges listed are the result of estimates of exposure uncertainties (usually a factor of two), methodological uncertainties that led to alternate evaluations of risk, or exposure, or. in some cases, statistical uncertainties associated with small numbers. In addition to some controversy over the individual Kl's. there was widespread disagreement as to which studies should ultimately be included in the determination of an overall KL for lung cancer. Table 2.-Estimates of KL From Proposed Rule and Final Determination Pro posal final fang* Henderson A Enlortine..... 0.0047 0.0047 Wail) ala).... .............. ...... 0.0033 0.0033 Finkelawrin....................... 0.067 0.048 Pe------- ------ ----------- - 0.0076 0.0076 Dement at al.................... 0.042 0042 Berry end Newhouse........ 0.0006 0.0006 Seidman el al__ __ _____ 0.068 0.045 Selikoff el al...................... 0x020 0.020 AriltuneIk: Mean....... 0.027 0.016 Geometic Mean........ 0.0113 0.01 Median_____ __ ___ ... 0.0138 0.0139 (0022-0.011) (0,0016-0.0086) (0.033-0.13) (0.0000-0.023) (0.23-0.21) (0-0.0006) (.023-06) (0.008-0.03) In its preliminary assessment, OSHA used the eight non-mining-and-milling studies to derive an overall estimate of KL of 0.01. As noted in the November proposal: Considering the industrial processes other than mining and milling. OSHA believes 0.01 to be a reasonable estimate of KL. It is the geometric mean and median of the Kl's derived from studies of asbestos manufacturing and insulation application processes. The geometric mean hod the advantage of minimizing the influence of outlying values and a Kt of 0.01 is approximately within one order of magnitude of all the estimates of KL. In sum, the KL of 0.01 is a best estimate which contains appropriate recognition of studies with higher and lower values of Kl. It should be noted however, that the uncertainties around this estimate of Kl are such that an appropriate estimate of Kl could lie between 0.003 and 0.03 (48 FR 51125]. The distinct nature of mining-milling data (and hence, the estimate of KL from these data] has been considered earlier. There is some evidence that risks in the asbestos mining-milling operations are lower than other industrial operations due to differences in fiber size. This differential was discussed by Nicholson (Ex. 3D3A). Thus, in determining the best overall value for KL for the final rule, the data from mining and milling processes . were not considered. OSHA still believes it to be valid to employ the same eight studies it used to derive the estimates for the November and April proposals. As discussed earlier, OSHA modified some of the values of KL for the final rule. Based upon these revised values, OSHA has determined that the best estimate of KL is 0.01. the same value derived for the proposals. The values given under the final estimate column in Table 2 have an arithmetic mean of 0.019 and a geometric mean of 0.0T. OSHA believes it has chosen reasonable estimates for the individual KL's and has been responsive to the comments made by participants in the hearing. In some cases. OSHA has lowered its original value of the estimate of KL in light of these comments or the addition of new data indicating such a change was warranted. The end result is that these small changes in individual values have little effect on the overall Kt value. This is most likely due to the Agency's choice of a reasonable KL for the proposal. Some scientists have suggested that some asbestos processes such as asbestos textile manufacturing, may pose a greater hazard than other processes. As noted earlier, while mining and milling appear to pose a lesser carcinogenic hazard than , manufacturing processes, when OSHA compared the potency factors for lung cancer (KJ among different studies of different processes, no consistent pattern of differential lung cancer risk by process emerged. Therefore, again, the choice of a midpoint unit risk for all industrial processes is a reasonable and justified choice. In sum, the KL of 0.01 is a best estimate which contains appropriate recognition of studies with higher and lower values of Kt. It should be noted, however, that the uncertainties around this estimate ofKL are such that an appropriate estimate of KL could lie between 0.003 and 0.3. As discussed earlier, Crump believed that both the Seidman et al. and Selikoff et al. studies should have been excluded from the calculation of KL. Along with the other adjustments discussed above, Crump estimated an overall Kl of 0.0085. As Dr. Crump noted in his testimony: OSHA has developed what I would term an upper limit assessment of asbestos riBk. In dealing with uncertainty, OSHA has. in a number of instances, made assumptions that tend to minimize the possibility of underestimating.the risk. In addition, the uncertainties in some of their assumptions appear to be underestimated by OSHA. The three most sighifleant assumptions in OSHA's risk assessment that lead to upper limit estimates of risk are the assumptions of: (l) a linear dose-response relationship: (2) the same potency for all forms of asbestos; and (3) attribution of the lung cancer component of risk caused by smoking to the overall risk of asbestos |Ex. 237A, p. 4-5]. However, in addition to Dr. Crump's recommendations, several commenters noted a number of different ways for incorporating the available data into an overall estimate of risk. For example, in his written testimony. Dr. Marvin Schneiderman, who served as a member of CHAP and who was one of the reviewers of OSHA's November GLEASON-000885