Table AIV shows the marginal distribution of person-years by facility and dose group for X-10/Y-12 workers (white males) with a ten year lag. Table AIV also shows the marginal average dose, i.e., the person-year weighted average of all the average doses from each cell in the full table. The person-year distributions in this table shows that the dose distribution at Y-12 was considerably different from that of X-10 and multiple workers. The person-years in the X-10 only category is greater than reported in previous studies [7,6] since individuals who were initially employed at X-10 and then at another Oak Ridge facility contributed person-years to the appropriate dose groups in the X-10 only category until ten years (the lag) after the date at which they became a multiple facility worker. In most dose groups the marginal average dose was slightly less than the mid-point of the interval used to define the category. In dose-response analyses the average dose in each cell was used.
In previous reports for X-10 only workers [15,16] different dose groups were defined and the interval mid-point was used for all groups except the highest group where the median value of all person-years in that interval was used. Table AIV also shows the marginal distribution of observed and ``expected'' deaths (using U.S. rates) by facility and dose groups for all cancer deaths, and the marginal total for each dose-group. These unadjusted SMRs indicate that mortality rates for all cancer causes are less than those for U.S. White males. To adjust for this selection effect we used the external standard rates and fit a multiplicative main effects model that includes all of the factors other than dose group. The `adjusted' expected deaths and SMRs are listed at the bottom of Table AIV. A GLIM4 program is provided with the all cancer data set that can be used to obtain Table AIV.