Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: test for homogeneous odds ratios in D+L random effects model


From   Xin Lu <[email protected]>
To   [email protected]
Subject   Re: st: test for homogeneous odds ratios in D+L random effects model
Date   Thu, 30 Sep 2010 10:27:31 -0500

The dependent variable in meta-regression would be the ln(OR), that
is, the pre-post variable is embedded in it already. The interaction I
would like to test is how the "pre-post" interacts with the "vendor
affiliation", and try to answer the question "whether vendor
affiliated studies would have a bigger impact on reducing mortality in
the "post" comparing to "pre" ".

Thanks,


On Thu, Sep 30, 2010 at 3:05 AM, Jonathan Sterne
<[email protected]> wrote:
> Hi - you need meta-regression implemented in the Stata command metareg - I
> recommend that you read Roger Harbord's recent article in the Stata Journal.
>
> Best wishes
>
> Jonathan Sterne
>
> --On 30 September 2010 02:33 -0400 statalist-digest
> <[email protected]> wrote:
>
>> Date: Wed, 29 Sep 2010 11:33:55 -0500
>> From: Xin Lu <[email protected]>
>> Subject: st: test for homogeneous odds ratios in D+L random effects model
>>
>> We are collaborating with some investigators off site on a
>> meta-analysis regarding mortality of tele-ICU?implementation (Post
>> versus Pre). We are using STATA metan package to conduct D+L random
>> effects model to pool the odds ratios of mortality from various
>> studies, and also conducting subgroup analysis based on vendor
>> affiliation (vendor affiliated versus not vendor affiliated). The
>> investigators off site asked us to test for homogeneous odds ratios
>> between the two subgroups (vendor affiliated versus not vendor
>> affiliated). However, as far as we can find, such test (breslow-day
>> test) can only be performed using fixed inverse variance model, which
>> differs from the D+L random effects model we are using. They insisted
>> there should be a way to do it. We are not expert on meta-analysis,
>> and therefore we can't say for sure there is no way to achieve it.
>>
>> Any insights?
>>
>> Thanks a million!
>> Xin
>>
>
>
>
>
> ----------------------
>
> Jonathan Sterne
> Professor of Medical Statistics and Epidemiology
> School of Social and Community Medicine
> University of Bristol
> Canynge Hall
> 39 Whatley Road
> Bristol BS8 2PS
> UK
>
> Tel:    0117 928 7396
> Fax:    0117 928 7325
> E-mail: [email protected]
> web:    www.epi.bris.ac.uk/staff/jsterne.htm
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index