CFR > Title 42 - Public Health > CHAPTER I—PUBLIC HEALTH SERVICE, DEPARTMENT OF HEALTH AND HUMAN SERVICES > SUBCHAPTER G—OCCUPATIONAL SAFETY AND HEALTH RESEARCH AND RELATED ACTIVITIES (parts 80 to 87) > PART 81—GUIDELINES FOR DETERMINING PROBABILITY OF CAUSATION UNDER THE ENERGY EMPLOYEES OCCUPATIONAL ILLNESS COMPENSATION PROGRAM ACT OF 2000 > SUBPART E—Guidelines To Estimate Probability of Causation (§81.20 to §81.30) > 81.25—Guidelines for claims including two or more primary cancers.

81.25—Guidelines for claims including two or more primary cancers.

For claims including two or more primary cancers, DOL will use NIOSH-IREP to calculate the estimated probability of causation for each cancer individually. Then DOL will perform the following calculation using the probability of causation estimates produced by NIOSH-IREP:

Code of Federal Regulations

Calculate: 1−[{1−PC1} × {1−PC2}×. . . × {1−PCn}] = PCtotal,
where PC1 is the probability of causation for one of the primary cancers identified in the claim, PC2 is the probability of causation for a second primary cancer identified in the claim, and PCn is the probability of causation for the nth primary cancer identified in the claim. PCtotal is the probability that at least one of the primary cancers (cancers 1 through “n”) was caused by the radiation dose estimated for the claim when Equation 1 is evaluated based on the joint distribution of PC1, . . ., PCn.4 DOL will use the probability of causation value calculated for PCtotal to adjudicate the claim.

Code of Federal Regulations

Footnote(s): 4 Evaluating Equation 1 based on the individual upper 99th percentiles of PC1, . . ., PCn approximates the upper 99th percentile of PCtotal whenever PC1, . . ., PCn are highly related, e.g., when a common dose-reconstruction is the only non-negligible source of uncertainty in the individual PCi's. However, this approximation can overestimate it if other sources of uncertainty contribute independently to the PC1, . . ., PCn, whereas treating the joint distribution as fully independent could substantially underestimate the upper 99th percentile of PCtotal whenever the individual PCi's are positively correlated.
[67 FR 22309, May 2, 2002; 67 FR 62096, Oct. 3, 2002]