I suggest you try the modified Poisson regression approach suggested by Zou (Am J Epidemiol 2004;159:702-6). This is the standard Poisson model, with the 'irr' option used to obtain the relative risk and the 'robust' variance estimator used to obtain an appropriate 95% CI. This will enable you to fit models with confounding variables.
Roger Webb
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Charlotta Eriksson
Sent: 24 October 2005 16:08
To: [email protected]
Subject: st: analysis of cumulative incidence data
Dear Statalist User,
I'm about to analyze a cohort of men regarding the cumulative incidence of
hypertension in relation to a categorized exposure variable. Currently I'm
considering which method that would be most suitable for calculating risk
estimates and I turn to you for advice. I have considered the 'logistic'
command but that merely yields Odds Ratios and not the Relative Risks I'm
after. Then I found the 'binreg' command and was happy to see that it
yielded the same result as my "by hand calculations" (with inclusion of the
option rr), but when I tried to control for more than two confounding
variables, the model could not give an answer.
my questions:
1. Can the logistic regression be used in the analysis of cumulative
incidence data? How?
2. If not; what command is better suited for this purpose?
3. Is the 'binreg' command just a sub group to 'logistic'?
4. How do I adjust for different variables using the 'binreg' command?
Most great full for answers!!
Greetings from Charlotta
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
