Re: st: Meta-analysis of AUC curves developed using different type of regression models

["Roger B. Newson" <r.newson@imperial.ac.uk>]

To meta-analyse AUCs, I would use the -somersd- package, downloadable 
from SSC, with the option -transf(c)- to specify confidence intervals 
for Harrell's c (an alias of the AUC). I would use the -parmest- 
package, also downloadable from SSC, to produce an output dataset (or 
resultsset) with 1 observation for each subpopulation AUC and data on 
its estimate, standard error and confidence interval. Once you have this 
resultsset, you have the choice of doing the meta-analysis using the 
-metan- command of Jonathan Sterne, Roger Harbord et al., or 
(alternatively) using the -metaparm- module of the -parmest- package, 
together with the -parmhet- package (also downloadable from SSC) to do 
heterogeneity tests.

I hopwe this helps.

Best wishes

Roger


Roger B Newson BSc MSc DPhil
Lecturer in Medical Statistics
Respiratory Epidemiology and Public Health Group
National Heart and Lung Institute
Imperial College London
Royal Brompton Campus
Room 33, Emmanuel Kaye Building
1B Manresa Road
London SW3 6LR
UNITED KINGDOM
Tel: +44 (0)20 7352 8121 ext 3381
Fax: +44 (0)20 7351 8322
Email: r.newson@imperial.ac.uk
Web page: http://www.imperial.ac.uk/nhli/r.newson/
Departmental Web page:
http://www1.imperial.ac.uk/medicine/about/divisions/nhli/respiration/popgenetics/reph/

Opinions expressed are those of the author, not of the institution.

On 19/07/2012 15:20, Etan Lakam wrote:
> Dear Listers,
>
> I have a series of studies of prediction  of an outcome , in which the
> area under the curve (AUC) or C-statistic was estimated using
> different variables and different types of regression models (Weibull
> and Logistic), and I would like to combine the AUCs in a
> meta-analysis.
>
> I have two questions:
>
> 1- Which Stata command I use to combine the AUCs if I have these
> estimates and their standard error or 95%CI confidence intervals?   I
> really can't get my head around this.
>
>
> 2- Is such a meta-analysis a sensible thing to do given the
> heterogeneity linked to  the differences in regression models (Weibull
> and logistic) used for model developement and the differences in
> variables used for prediction across studies and?
>
>
> Many thanks,
>
>
> Etan
>
> PS: Any documentation on the topic (metanalysis of AUCs curves) will
> also be welcome.
> *
> *   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/