thanks; forgot about mhodds.
John LeBlanc
On Thu, 19 Apr 2007 17:12:58 -0400, Visintainer, Paul wrote:
You can also use the -mhodds- command (in long format) to reproduce the
OR and 95% CI for pair-matched data (1:1 match). Reshaping the data is
not necessary. For instance, you can use:
mhodds case exp pair
- where pair is the ID for the pair;
For more than one control per case or irregular matching, -clogit- is
probably the best choice.
-p
______________________________________
Paul F. Visintainer, PhD
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Svend Juul
Sent: Thursday, April 19, 2007 8:24 AM
To: [email protected]
Subject: Re: st: Matched case-control studies: useful Stata commands
esp. for description
From John LeBlanc
To
Subject st: Matched case-control studies: useful Stata commands esp.
for description
Date Wed, 18 Apr 2007 16:33:06 -0300
------------------------------------------------------------------------
--------
John LeBlanc wrote:
Could someone give me some general advice on useful Stata commands for
analyzing matched case-control studies? I have study with a variable #
of controls/case arranged longitudinally. The outcome variable is a road
traffic injury to a child and I have ~50 cases and 4-5 controls/case.
Here's what I have gleaned so far from reading the Stata manuals,
searching Google and this archive.
Multivariate analysis and hypothesis testing
clogit: seems to be the single most useful command since it can generate
measures of association, C.I. & p-values for variable number of cases &
controls
1:1 matched study
epitab mcc
BUT, data must be wide and only for case-control pairs.
DESCRIPTIVE STATISTICS
This is where I'm having the most trouble. I"d like to be able to
summarize various demographic characteristics and other independent
variables for cases and controls before entering them into a
multivariate clogit model.
Do I just use standard commands such as "table" or "means" and ignore
any tests of hypotheses since I would get these from Stata?
Are there any Stata commands that are particularly useful for
summarizing a matched case-control dataset?
Do I simply ignore the matching in the descriptive analysis and present,
e.g., means or frequency distributions for the cases and controls by the
various independent variables?
------------------------------------------------------------
You are right about the use of the -clogit- and -mcc- commands for
matched case-control data.
Your question about descriptive statistics is less clear. Are you
asking:
"What to put in Table 1?"? The readers (or a reviewer) might be
interested
in assurance that the matching was successful, or they may want to see
general characteristics of the study population. I don't see that this
requires other than standard descriptive procedures.
Hope this helps
Svend
__________________________________________
Svend Juul
Institut for Folkesundhed, Afdeling for Epidemiologi
(Institute of Public Health, Department of Epidemiology)
Vennelyst Boulevard 6
DK-8000 Aarhus C, Denmark
Phone: +45 8942 6090
Home: +45 8693 7796
Email: [email protected]
__________________________________________
*
* 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/
*
* 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/
