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: New on SSC: ipdmetan


From   David Fisher <[email protected]>
To   [email protected]
Subject   Re: st: New on SSC: ipdmetan
Date   Tue, 11 Feb 2014 10:02:13 +0000

Dear all,

With thanks (and apologies) to Kit Baum, a minor update to -ipdmetan-
is now available from SSC.

A bug has been fixed that would cause the program to exit with error
r(3499) "Mata error" unless specific Mata routines were installed.
The program now exhibits the correct behaviour, which is to run
successfully if the Mata routines are not required, or to exit with a
helpful error message if they *are* required (for some random-effects
models only).

Many thanks,

David.

David Fisher
Statistician
MRC Clinical Trials Unit at UCL
e-mail: [email protected]


P.S.  This is a useful demonstration of the difference between Mata
and Stata.  My original code included calls to the routines "mm_root"
and "integrate" which are necessary for particular random-effects
models, within if{} statements relevant to those models.  I also
included error-traps near the start of my Stata code to check which
random-effects model was specified; and if it required a particular
Mata routine, whether that routine was installed.

However, as Mata code is compiled immediately, those calls to
"mm_root" etc. are noticed as soon as Mata is called from within
-ipdmetan-, and the program exits with error *regardless* of what
random-effects model was specified. The solution was to move all calls
to "mm_root" etc. into Mata subroutines, so that the main routine
directly called from within -ipdmetan- contained standard Mata
commands only.





On Mon, Feb 3, 2014 at 7:23 PM, David Fisher <[email protected]> wrote:
> Dear all,
>
> A new meta-analysis package, -ipdmetan-, is available from SSC with thanks
> to Kit Baum.
>
> The main program carries out two-stage inverse-variance IPD meta-analysis
> by looping over a series of categories, fitting the desired model to the
> data within each.  The screen output and forestplot are based on those in
> the existing program -metan-, with thanks to Ross Harris and others.
>
> A variety of analyses are possible: a greater variety of random-effects
> models are available; aggregate data may be included from an external
> dataset; and a separate program -ipdover- may be used to create a
> forestplot of a series of (potentially overlapping) subgroups within a
> single study or trial.  The forestplot routine from -metan- has been
> updated and can now be run by itself (program -forestplot-).  It now has
> greater flexibility in terms of rendering; and titles, formatting etc may
> be tweaked by making changes to the dataset itself beforehand.
>
> Help files are available, but no examples as yet, sorry!  Please contact me
> if you have any questions, problems or suggestions.
>
> Many thanks,
>
> David.
>
>
> David Fisher
> Statistician
> MRC Clinical Trials Unit at UCL
> e-mail:  [email protected]
> *
> *   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/


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