Skip to main content
. 2016 Jun 10;125(2):223–229. doi: 10.1289/EHP151

Table 3.

Breast cancer incidence risk by sex, using stratified relative risk Model 1 with strata defined by city, sex, age at exposure, attained age, using data of Preston et al. (2007).

Model number Model/parameter fitted ERR/Sv (95% CI) Temporal modifiers (95% CI) Deviance p-Valuea
1 Male breast cancer (αm) 19.41 (1.53, 761.30) 53.95 0.003b
2 Female breast cancer (αf) 1.50 (1.12, 1.95) 2770.49 < 0.001b
3 Male and female breast cancer (α) 1.54 (1.15, 1.98) 2828.32 < 0.001b
0.049c
4 Adjusted for attained age (α) 1.90 (1.41, 2.50) 2820.14 0.004
Attained age adjustment (per 10 years of age) (exp[β1]) 0.70 (0.52, 0.90)
5 Adjusted for time since exposure (α) 1.62 (1.20, 2.13) 2827.37 0.331d
Time since exposure adjustment (per 10 years of time since exposure) (exp[β2]) 0.89 (0.70, 1.13)
6 Adjusted for age at exposure (α) 1.59 (1.19, 2.05) 2825.28 0.081d
Age at exposure adjustment (per 10 years of age at exposure) (exp[β3]) 0.82 (0.64, 1.02)
7 Adjusted for attained age and age at exposure (α) 1.89 (1.39, 2.49) 2820.07 0.016d
Attained age adjustment (per 10 years of age) (exp[β1]) 0.71 (0.52, 0.95) 0.023e
Age at exposure adjustment (per 10 years of age at exposure) (exp[β3]) 0.97 (0.73, 1.26) 0.802f
8 Adjusted for sex, attained age and age at exposure: males (αm) 27.68 (1.81, 90.16g) 2816.24 0.050
Adjusted for sex, attained age and age at exposure: females (αf) 1.86 (1.36, 2.46)
Attained age adjustment (per 10 years of age) (exp[β1]) 0.70 (0.51, 0.94)
Age at exposure adjustment (per 10 years of age at exposure) (exp[β3]) 0.99 (0.75, 1.30)
Note: ––, data not available. aUnless otherwise indicated the p-value represents the improvement in fit over the model in the row immediately above it; all model numbers refer to the leftmost column. bp-Value of improvement in fit over the null model, with no terms in dose. cp-Value of improvement in fit of model allowing for different dose coefficients (αm, αf) by sex, over Model 3, in other words what is obtained by comparing the combined fit of Models 1 and 2 with that of Model 3. dp-Value of improvement in fit over Model 3. ep-Value of improvement in fit over Model 6. fp-Value of improvement in fit over Model 4. gWald-based CI.