Skip to main content
. 2014 Aug 11;14:584. doi: 10.1186/1471-2407-14-584

Table 3.

Multivariable model for older group using stepwise regression with AIC criterion*

Risk factor Beta Odds ratio 95% CI (Lower -Upper) p value
(Intercept) −1.16 0.31 0.18 - 0.55 0.000 ***
Palpable lump 0.013 **
No corresponding palpable mass 0.00 1(referent)
Missing −0.30 0.74 0.05 - 10.55 0.824
Corresponding palpable mass 0.80 2.22 1.12 - 4.41 0.022 **
Family history 0.043 **
None 0.00 1(referent)
Missing −0.89 0.41 0.13 - 1.32 0.135
Strong −0.32 0.73 0.33 - 1.59 0.422
Very strong 1.66 5.24 0.84 - 32.78 0.076 *
Prior surgery 0.132
Not present 0.00 1(referent)
Missing −0.36 0.70 0.07 - 6.82 0.759
Present 0.57 1.78 0.99 - 3.17 0.053 *
Principal mammography finding <0.001 ***
Calcifications or Single dilated duct 0.00 1(referent)
Architectural distortion 20.56 Inf 0.00 - Inf 0.993
Associated calcifications 2.16 8.67 3.39 - 22.14 0.000 ***
Missing 2.10 8.14 3.88 - 17.09 0.000 ***
Asymmetry or Focal asymmetry 2.94 18.87 3.79 - 93.87 0.000 ***
Mass 3.04 20.93 9.20 - 47.65 0.000 ***
Developing asymmetry 2.80 16.45 1.78 - 151.95 0.014 **
Calcification distribution 0.008 **
Not present 0.00 1(referent)
Linear or Segmental −3.11 0.04 0.00 - 0.49 0.011 **
Clustered −0.69 0.50 0.22 - 1.18 0.113
Regional or Scattered −1.94 0.14 0.01 - 2.83 0.202
Mass margins <0.001 ***
None 0.00 1(referent)
Circumscribed −2.51 0.08 0.01 - 0.45 0.004 ***
Ill-Defined 0.19 1.21 0.46 - 3.20 0.703
Obscured 0.10 1.10 0.11 - 11.31 0.935
Spiculated 28.70 Inf 0.00 - Inf 0.983
Mass shape 0.033 **
None 0.00 1(referent)
Irregular 1.91 6.78 0.78 - 58.79 0.083 *
Lobular or Oval −0.13 0.87 0.24 - 3.16 0.838
Round −15.53 0.00 0.00 - Inf 0.987
Focal asymmetric density 0.077 *
Not present 0.00 1(referent)
Present 1.63 5.10 0.54 - 47.77 0.154

The model is presented in the order of inclusion into the model.

*Asterisks denote the level of significance such that: ***p-value < 0.001; **p-value < 0.05, and *p-value <0.1.

“Inf” (short for infinity) is inserted at places where the data for the corresponding variable is sparsely populated and produces a very high and unstable odds ratio.