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

Test The following questions and computer output consider a
data from a cross-sectional study carried out at Grady
Hospital in Atlanta, Georgia involving 289 adult patients
seen in an emergency department whose blood cultures
taken within 24 hours of admission were found to have
Staph aureus infection (Rezende et al., 2002). Information
was obtained on several variables, some of which were
considered risk factors for methicillin resistance (MRSA).
The outcome variable is MRSA status (1¼yes, 0¼no),
and covariates of interest included the following vari-
ables: PREVHOSP (1¼previous hospitalization, 0¼no
previous hospitalization), AGE (continuous), GENDER
(1¼male, 0¼female), and PAMU (1¼antimicrobial
drug use in the previous 3 months, 0¼no previous anti-
microbial drug use).

The SAS output provided below was obtained for the fol-

lowing logistic model:

Logit PðXÞ¼aþb 1 PREVHOSPþb 2 AGEþb 3 GENDER

þb 4 PAMU

Analysis of maximum likelihood estimates

Wald

Parameter DF Estimate

Standard

Error

Chi-

Square Pr>ChiSq

Intercept 1 5.0583 0.7643 43.8059 <.0001

PREVHOSP 1 1.4855 0.4032 13.5745 0.0002

AGE 1 0.0353 0.00920 14.7004 0.0001

gender 1 0.9329 0.3418 7.4513 0.0063

pamu 1 1.7819 0.3707 23.1113 <.0001

Odds Ratio Estimates

Point 95% Wald

Effect Estimate Confidence Limits

PREVHOSP 4.417 2.004 9.734

AGE 1.036 1.017 1.055

gender 2.542 1.301 4.967

pamu 5.941 2.873 12.285

Association of Predicted Probabilities and Ob-

served Responses

Percent Concordant 83.8 Somers' D 0.681

Percent Discordant 15.8 Gamma 0.684

Percent Tied 0.4 Tau-a 0.326

Pairs 19950 c 0.840

380 10. Assessing Discriminatory Performance of a Binary Logistic Model