Logistic Regression: A Self-learning Text, Third Edition (Statistics in the Health Sciences)

B. Example of model involving three exposure

variables:

logit PðÞ¼X aþb 1 SMKþb 2 PALþb 3 SBP

þ~

p 1

i¼ 1

giVi:

C. The odds ratio formula for the general no

interaction model:

RORE*vs:E**¼exp½ðE* 1 E** 1 Þb 1 þðE* 2 E** 2 Þb 2

þþðE*qE**qÞbqŠ;

whereE*¼(E 1 *,E 2 *,...,Eq*) andE**¼

(E 1 *,E 2 **,...,Eq**) are two specifications of the

collection of exposure variables to be compared.

D. Example of odds ratio involving three exposure

variables:

RORE*vs:E**¼exp½ðSMK*SMK**Þb 1

þðPAL*PAL**Þb 2

þðSBP*SBP**Þb 3 Š:

VI. The model and odds ratio for several exposure

variables with confounders and interaction

(pages 87–91)

A. An example of a model with three exposure

variables:

logit PðÞ¼X aþb 1 SMKþb 2 PALþb 3 SBPþg 1 AGE

þg 2 SEXþSMKðÞd 11 AGEþd 12 SEX

þPALðÞd 21 AGEþd 22 SEX

þSBPðÞd 31 AGEþd 32 SEX:

B. The odds ratio formula for the above model:

RORE*vs:E**¼exp½ðSMK*SMK**Þb 1

þðPAL*PAL**Þb 2 þðSBP*SBP**Þb 3

þd 11 ðSMK*SMK**ÞAGE

þd 12 ðSMK*SMK**ÞSEX

þd 21 ðPAL*PAL**ÞAGE

þd 22 ðPAL*PAL**ÞSEX

þd 31 ðSBP*SBP**ÞAGE

þd 32 ðSBP*SBP**ÞSEXŠ

94 3. Computing the Odds Ratio in Logistic Regression