Skip to main content
. 2011 Nov 3;105(11):1669–1675. doi: 10.1038/bjc.2011.466

Table 3. Multivariate model for malignancy. A multivariate log binomial regression model including all predictors was fitted to the data.

Variable Relative risk Lower 95% CI Upper 95% CI P-value
Strong family history: yes vs no 1.12 0.99 1.28 0.08
         
Age group (years)       0.14
 50–59 vs <50 1.03 0.89 1.18 0.72
 60–69 vs <50 0.96 0.82 1.11 0.58
 70+ vs <50 1.11 0.93 1.32 0.24
 60–69 vs 50–59 0.93 0.85 1.02 0.14
 70+ vs 50–59 1.08 0.95 1.22 0.23
 70+ vs 60–69 1.16 1.02 1.32 0.03
         
Screening round       <0.01
 2–5 vs 1 1.42 1.27 1.58 <0.01
 6+ vs 1 1.54 1.35 1.77 <0.01
 6+ vs 2–5 1.09 0.98 1.21 0.11
         
Palpable mass: yes vs no 1.13 1.03 1.24 0.01
         
Radiological grade       <0.01
 4 vs 3 2.20 1.94 2.51 <0.01
 5 vs 3 3.26 2.90 3.65 <0.01
 5 vs 4 1.48 1.36 1.61 <0.01
         
Distribution       0.52
 Multiple vs single 0.96 0.84 1.10 0.57
 Dispersed vs single 1.03 0.94 1.13 0.51
 Dispersed vs multiple 1.07 0.95 1.22 0.28
         
Size >15 vs 15 or less 1.29 1.17 1.44 <0.01