Table 2.

Values for ERR and EAR for cerebrovascular disease calculated with the multi-model inference, the ERR-LNT model, the ERR-quadratic model, and the ERR-step model for 0.2 and 1 Gy and different values of age at exposure (e) and age attained (a)

	ERR	EAR [per 104 PY]
CVD
Multi-model inference
0.2 Gy
e = 20, a = 50	0.007 (0, 0.035)	0.05 (0, 0.23)
e = 20, a = 70	0.007 (0, 0.035)	0.17 (0, 0.84)
e = 30, a = 70	0.007 (0, 0.035)	0.3 (0, 1.4)
e = 50, a = 70	0.007 (0, 0.035)	0.8 (0, 3.8)
1 Gy
e = 20, a = 50	0.165 (0.033, 0.32)	1.10 (0.22, 2.1)
e = 20, a = 70	0.165 (0.033, 0.32)	3.97 (0.78, 7.7)
e = 30, a = 70	0.165 (0.033, 0.32)	6.6 (1.3, 13)
e = 50, a = 70	0.165 (0.033, 0.32)	18.0 (3.6, 35)
Single models
0.2 Gy, e = 30, a = 70
ERR-LNT model [#1]	0.0248 (0.0055, 0.044)	0.98 (0.22, 1.7)
ERR-quadratic model [#2]	2.84 × 10−3 (4.0 × 10−4, 5.3 × 10−3)	0.114 (0.016, 0.21)
ERR-step model [#6], D th = 0.62 Gy	0	0
1 Gy, e = 30, a = 70
ERR-LNT model [#1]	0.124 (0.028, 0.22)	4.9 (1.1, 8.7)
ERR-quadratic model [#2]	0.071 (0.010, 0.13)	2.85 (0.40, 5.3)
ERR-step model [#6], D th = 0.62 Gy	0.22 (0.093, 0.34)	8.7 (3.7, 14)
Preston ERR-LNT model (Preston et al. 2003)	0.12 (0.02, 0.22)	5.0a (1.0, 8.9)
ERR-LNT model (Shimizu et al. 2010)	0.09 (0.01, 0.17)b	2.3 (0.4, 4.4)b

The 90% confidence intervals are provided. The risk values from Preston et al. (2003) and Shimizu et al. (2010) are also shown. The numbers in brackets refer to the eleven dose–responses depicted in Fig. 1. The EAR-values for MMI and for the single models #1, #2, and #6 are only valid for men in Hiroshima. The city-averaged EAR-values for men can be calculated by multiplication with a factor of 1.1 (see Sect. 6 of the Online Resource for mathematical details). The EAR-values for women can be calculated by multiplying with a factor of 0.6

^aNot given by Preston et al. (2003); calculated from Preston’s ERR-LNT model

^bThis is the 95% CI