Re: st: meta-analysis graphs with metan

[Roger Newson <roger.newson@kcl.ac.uk>]

At 18:23 22/03/2005, Michael McCulloch wrote:
> Metan is a very useful routine for calculating summary effect measures. 
> However, I am having a problem with balancing text size of study ID 
> (namevar) with the width of the text columns where event counts are 
> displayed. Here's the problem:
> 1. If I set texts to display at full size texts=1, then the text for the 
> event counts in group 1 overlaps with that of group 2.
> 2. If I set texts to display at reduced size (texts=0.7), then the event 
> counts text columns are properly separated, but the study ID text 
> (namevar) is so small it is unsuitable for publication.
>
> I should add that the # of events is large, four digits, making the 
> display issue more problematic. Techniques for customizing the graph are 
> only minimally documented in "help metan."
>
> I'd greatly appreciate any suggestions on how to customize the graph. 
> Especially helpful would be suggestions on how to increase the width 
> between the event counts text columns.
I don't have much experience with -metan-. However, if you want the full 
power of Stata 8 graphics (including the -ysize()- option), then you might 
like to use the -eclplot- package, downloadable from SSC, to create forest 
plots. -eclplot- inputs a dataset (or resultsset) with 1 observation per 
confidence interval to be displayed, and data on estimates, confidence 
limits and (optionally) study sizes, and produces as output a confidence 
interval plot (optionally horizontal like a Cochrane forest plot), with the 
option of weighting the estimate symbols (or even the confidence limit 
caps) by study size. The plot can be customized using the -plot()- option 
and other Stata 8 graphics options.

To create the input dataset in the first place, a useful tool might be the 
-metaparm- package, also downloadable from SSC. -metaparm- takes, as input, 
a dataset with 1 observation per study and data on estimates, standard 
errors and (optionally) by-groups, and creates, as output, an output 
dataset (or resultsset) with one observation, or one observation per 
by-group, and data on estimates, confidence limits and numbers of subjects 
for the overall summary parameter. This resultsset can be -append-ed to the 
original dataset to produce a dataset ready to input to -eclplot- to 
produce the forest plot.

I hope this helps.

Roger



--
Roger Newson
Lecturer in Medical Statistics
Department of Public Health Sciences
Division of Asthma, Allergy and Lung Biology
King's College London

5th Floor, Capital House
42 Weston Street
London SE1 3QD
United Kingdom

Tel: 020 7848 6648 International +44 20 7848 6648
Fax: 020 7848 6620 International +44 20 7848 6620
  or 020 7848 6605 International +44 20 7848 6605
Email: roger.newson@kcl.ac.uk
Website: http://phs.kcl.ac.uk/rogernewson/

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

*
*   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/