[R] questions about meta-analysis

[Emmanuel Charpentier]

Le samedi 27 juin 2009 à 13:02 +0800, sdzhangping a écrit :
> Dear R users:
> 
> In the example of meta-analysis (cochrane, package rmeta), I can not
> found the p-value of Test for overall effect, and some other indices
> (Z, I, weight and et al).  How can I get the these indices listed? 
> > library(rmeta)
> > data(cochrane)
> > cochrane
>           name ev.trt n.trt ev.ctrl n.ctrl
> 1     Auckland     36   532      60    538
> 2        Block      1    69       5     61
> 3        Doran      4    81      11     63
> 4        Gamsu     14   131      20    137
> 5     Morrison      3    67       7     59
> 6 Papageorgiou      1    71       7     75
> 7      Tauesch      8    56      10     71
> > a=meta.MH(n.trt,n.ctrl,ev.trt,ev.ctrl,names=name,data=cochrane)
> > summary(a)
> Fixed effects ( Mantel-Haenszel ) meta-analysis
> Call: meta.MH(ntrt = n.trt, nctrl = n.ctrl, ptrt = ev.trt, pctrl =
> ev.ctrl, 
>     names = name, data = cochrane)
> ------------------------------------
>                OR (lower  95% upper)
> Auckland     0.58    0.38       0.89
> Block        0.16    0.02       1.45
> Doran        0.25    0.07       0.81
> Gamsu        0.70    0.34       1.45
> Morrison     0.35    0.09       1.41
> Papageorgiou 0.14    0.02       1.16
> Tauesch      1.02    0.37       2.77
> ------------------------------------
> Mantel-Haenszel OR =0.53 95% CI ( 0.39,0.73 ) (where is Z and
> p-value ?)
> Test for heterogeneity: X^2( 6 ) = 6.9 ( p-value 0.3303 )

You might easily recompute them : use the confidence interval to
guesstimate the standard deviation of OR, its value and the damn Z and p
you long for... Remember that it's log(OR) that is (asymptotically)
normally distributed. The weights are proportional to inverse of
variance and sum to 1. You might peek at the source code for
enlightment...

Alternatively, you may grow lazy an use the "meta" package, whose
function metabin() will give you all that. The recent "metafor" package
seems also quite interesting.

Another (smarter(?)) alternative is to ask yourself *why* you need Z and
p. Z is just a computing device allowing you to use a known density. And
the real meaning of p and its usefulness has been discussed, disputed
and fought over at exceedingly large lengths in the past 80 years.

Of course, if you have an instructor to please ...

HTH,

					Emmanuel Charpentier