st: RE: Combining catagories of risk

[Timothy Mak <tshmak@hku.hk>]

Of course not! 
-metan- is for meta-analysis, where the different units are supposed to be independent. Analyses from the same studies are generally not independent because they use the same data. Furthermore, even if they are independent, there's no rationale for taking the average (which is what -metan- does) as an estimate for the HR of 'any renal dysfunction'. 

Having said that, sometimes in a meta-analysis, you want to extract just one estimate from a study, when the study presents many similar results. You may want to use the average of the different results, but if you use -metan-, that will underestimate the standard error. Perhaps a more reasonable strategy is to average the variance or even the SE from the different estimates. But this is just an ad hoc idea in the case where no better strategy exists. A more usual strategy is to find the estimate that is most relevant to your meta-analysis. 

Tim

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Claire Rushton
Sent: 26 March 2014 21:40
To: statalist@hsphsun2.harvard.edu
Subject: st: Combining catagories of risk

A study has produced risk estimates by different catagories of renal
dysfunction. To produce an overall risk estimate of 'any renal
dysfunction' is it appropriate to combine the risk estimates from the
different categories of renal dysfunction? (The estimates are hazard
ratios).

If it is appropriate could METAN be used for this?

Thank you
Claire
--
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/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/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/