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

Detailed Outline

Abbreviated Outline

I. Overview (page 244)

Focus: Five issues not considered in Chaps. 6 and 7

 Apply to any regression analysis but focus on

binary logistic model

 Goal: determine “best” model

1. Modeling strategy when there are two or more
   exposure variables


2. Screening variables when modeling


3. Collinearity diagnostics


4. Influential observations


5. Multiple testing
   II. Modeling Strategy for Several Exposure Variables
   (pages 244–262)
   A. Extend modeling strategy for (0,1) outcome,k
   exposures (Es), andpcontrol variables (Cs)
   B. Example with twoEs: Cross-sectional study,
   Grady Hospital, Atlanta, GA, 297 adult patients
   Diagnosis: Staphylococcus aureus Infection

PREVHOSP, PAMU?

controlling for AGE, GENDER

MRSA,

Question:

C. Modeling strategy summary: SeveralEs andCs

Model: Logit PðXÞ

¼aþ~

q

i¼ 1

biEiþ~

p 1

j¼ 1

gjVjþ~

q

i¼ 1

~

p 2

k¼ 1

dikEiWk

þ~

q

i¼ 1

~

q

i^0 ¼ 1

i 6 ¼i^0

d*ii 0 EiEi^0

Step 1: Define initial model (above formula)

Step 2: Assess interaction

Option A: Overall chunk test

þOptions B or C

Option B: TestEWs, thenEEs

Option C: TestEWs, but assessVs before

EEs

Step 3: Assess confounding and precision (Vs)

Options A and B (continued):Vs after

EWs andEEs

Options C (continued):Vs afterEWs,

but prior toEEs

Step 4: Test for nonsignifEs if not components of

significantEEs

286 8. Additional Modeling Strategy Issues