Cytel's LogXact 5 vs. SAS's PROC LOGISTIC:
Urinary Tract Infection and Contraceptive Use

(Source: LogXact 5 user guide, from data gathered at the University of Michigan's Dept. of Epidemiology.)How is the development of first-time urinary tract infection (UTI) related to contraceptive use? A study of sexually active college women with UTI was conducted, and their use of various contraceptives was surveyed.
(437 observations, 9 covariates.) Covariates:

    AGE: 1=age>=24 
     OC: 1 = oral contraceptive 
CURRENT: 1 = no regular current sex partner 
 PASTYR: 1 = no regular partner with relationship < 1yr 
     VI: 1 = vaginal intercourse 
    VIC: 1 = vaginal intercourse with condom 
   VICL: 1 = vaginal intercourse with lubricated condom 
    VIS: 1 = vaginal intercourse with spermicide 
    DIA: 1 = diaphragm use

The results are displayed below. Of note is the role played by the covariate for diaphragm use-DIA. The primary question is whether the use of diaphragm is risky in the presence of other confounders Standard maximum likelihood logistic regression will not produce a model fit if this covariate is included, and yet this is an important covariate -- all seven diaphragm users developed UTI. Exact logistic regression is required for fitting this model, and SAS's PROC LOGISTIC cannot do it. Only LOGXACT 5 can solve the problem.

LOGXACT 5 RESULTS

=============================================================================================
Parameter Estimates
=============================================================================================
                   Point Estimate            Confidence interval and P-value for Beta
         Type  Beta        SE(Beta)    Type        95.0%     C.I.            Pvalue   SE    
                                                   Lower       Upper      2*1-sided   
=============================================================================================
age       MLE  ?           ?           Asymptotic  ?           ?          ?                  
current   MLE  ?           ?           Asymptotic  ?           ?          ?
                   
dia       MLE  ?           ?           Asymptotic  ?           ?          ?             
          MUE  2.0500      NA          Monte Carlo -0.0726     +INF       0.0564   0.0033    
                                       ( Seed=1035236606,Samples=10000 )

oc        MLE  ?           ?           Asymptotic  ?           ?         ?                   
pastyr    MLE  ?           ?           Asymptotic  ?           ?         ?                   
vi        MLE  ?           ?           Asymptotic  ?           ?         ?                   
vic       MLE  ?           ?           Asymptotic  ?           ?         ?                   
vicl      MLE  ?           ?           Asymptotic  ?           ?         ?                   
vis       MLE  ?           ?           Asymptotic  ?           ?         ?                   
%CONST    MLE  ?           ?           Asymptotic  ?           ?         ?                   
=============================================================================================

Analysis time :00:10:19
=============================================================================================

SAS's PROC LOGISTIC RESULTS
WARNING: There is not enough memory available for exact computations.

Try it yourself!
The links below include data files and SAS code that you can download.

LOGXACT 5--
Download LogXact data
View Instructions for LogXact-5 analysis

PROC LOGISTIC--
Download SAS code (includes data)

---

Back