Table 4.

Parameter estimates of the excess relative rate (ERR) and excess absolute rate (EAR) models for solid cancer and 5-year lagged colon dose in subsets of the life span study and INWORKS

	Life Span Study					INWORKS
	Dose coefficient at 1 Gy	90% CI	Deviance	df	P1df	Dose coefficient at 1 Gy	90% CI	Deviance	df	P1df
ERR modela
ERR(d) = β1d			12,404.9	29,058				5541.4	11,006
β1: linear	0.28	0.18; 0.38				0.29	0.07; 0.53
ERR(d) = β1d + β2d2			12,404.6	29,057	0.548c			5541.3	11,005	0.909c
β1: linear	0.23	0.10; 0.40				0.27	– 0.14; 0.68
β2: quadratic	0.03	– 0.05; 0.10				0.06	– 0.78; 1.04
ERR(d,a) = β1d exp(υa)			12,401.1	29,056	0.147d			5,539.0	11,004	0.307d
ERR/Gy at attained age < 60 years	0.35	0.19; 0.57				0.35	– 0.26; 1.04
ERR/Gy at attained age 60- < 80 years	0.31	0.20; 0.43				0.19	– 0.05; 0.46
ERR/Gy at attained age 80+ years	0.16	0.06; 0.29				0.86	0.20; 1.61
EAR modelb
EAR(d) = β1d			12,831.6	29,395				6329.2	11,848
β1: linear	8.03	3.74; 13.07				1.68	< 0; 7.55
EAR(d,a) = β1d exp(υa)			12,812.4	29,393	< 0.001d			6322.1	11,846	0.029d
EAR/Gy at attained age < 60 years	5.84	3.07; 9.50				0.42	< 0; 6.10
EAR/Gy at attained age 60- < 80 years	20.79	11.56; 31.55				13.86	– 4.34; 33.16
EAR/Gy at attained age 80+ years	33.27	13.85; 60.23				190.40	58.67; 334.9

CI likelihood-based confidence interval

^aDefined as λ₀(c,s,b,a)[1 + ERR(d,a)], where d is colon dose (cumulative in INWORKS), c is city for the LSS and country for INWORKS, s is sex, b is birth year, a is attained age in three categories (< 60, 60–80, 80+ years)

^bDefined as λ₀(c,s,b,a) + EAR(d,a), the dose coefficients in the EAR model describe the excess cases per 10,000 person-years per 1 Gy

The LSS ERR and EAR estimates are weighted averages over sex with a weight of 0.88 for males and 0.12 for females

P_1df: p value of a likelihood ratio test vs. ^cβ₂ = 0 or ^dυ = 0

 < 0: lower CI bound not estimated (on the boundary of the parameter space)