The risk coefficient, E, for different age groups, computed from total dose resulting from internal plus external exposure for Marshall Islanders, ranged from 1.5x10 to 7.4x10 per person-rad-years at risk, assuming a 1l0- year latent period, and 1.1xl0° to 5.8x10°, asssuming a 5-year latent period. These data are indicated in Tables 5 and 6, respectively. The total risk coefficient, E, was used in Eq. (1) to determine the internal risk coefficient, B. For external risk coeffic ents and lO-year latent period, we chose 3.3xl0°” for age <20 and 1.0xl0 for age 220 based on the Working Grou study (8); for S-year latent period we chose 2.5xi0 for age <18 and 1.3x10 for age >18, based on NCRP Report 80 (2). The results for internal risk coefficients are in Table 7. Finally, as we explained in the Methods, we chose a special value for the <lO-year age group, since it was baged on a large group of children exposed to x rays (1). This value was 4.9x10 cancers per personrad-years at risk, and the estimate for the internal risk coefficient was 1.3xi0°°, virtually the same as the value given in Table 7 for the 10-year latent period. A tabulation of risk coefficient versus internal thyroid dose is given in Table 8. These internal dose groupings resulted in little variation in external dose as a function of age. These groupings were made to examine the affect of dose on the value for internal risk coeffictent. Table 8 Average Dose Versus Internal and External Risk Coefficients, 10-Year Latent Pertod Average Average Internal Internal External External Total Group Thyroid Risk Thyroid Risk Risk Age 1954 Dose, rad Coefficient? Dose, rad Coefficient? Coeffictent? 1400 1.4x107° 63 3.3x107° 1.5x107® <10 10-18 >18 560 400 8.0x1076 6.1x1076 78 66 3.3x107° 1.0x107° 7.4x107° 5.4x10~° 4This study. DReference 8. A sensitivity analysis, of the parameters in Eq. (1), shows that the value for the total risk coefficient, E, impacts greatly on the estimate of the internal risk coefficient, 3, in this specific Marshall Islands study. This is because of the wide difference between internal thyroid dose, A, and external thyroid dose, C. Thus, our estimate of internal risk coefficient depends largely on the observed tncidence of thyroid cancer because the total risk coefficient, E, is very sensitive to the small number of spontaneous and excess thyroid cancers observed. 27