metamerge  R Documentation 
This function can be used to merge pooled results of two metaanalyses into a single metaanalysis object. This is, for example, useful to produce a forest plot of a randomeffects metaanalysis with different estimates of the betweenstudy variance τ^2.
metamerge( meta1, meta2, pooled1, pooled2, text.pooled1, text.pooled2, text.w.pooled1, text.w.pooled2, label1, label2, backtransf )
meta1 
First metaanalysis object (see Details). 
meta2 
Second metaanalysis object (see Details). 
pooled1 
A character string indicating whether results of
common effect or random effects model should be considered for
first metaanalysis. Either 
pooled2 
A character string indicating whether results of
common effect or random effects model should be considered for
second metaanalysis. Either 
text.pooled1 
A character string used in printouts and forest plot to label the results from the first metaanalysis. 
text.pooled2 
A character string used in printouts and forest plot to label the results from the second metaanalysis. 
text.w.pooled1 
A character string used to label weights of the first metaanalysis. 
text.w.pooled2 
A character string used to label weights of the second metaanalysis. 
label1 
A character string used to label estimate of betweenstudy variance and heterogeneity statistics of the first metaanalysis. 
label2 
A character string used to label estimate of betweenstudy variance and heterogeneity statistics of the second metaanalysis. 
backtransf 
A logical indicating whether results should be
back transformed in printouts and plots. If

In R package meta, objects of class "meta"
contain
results of both a common effect and random effects
metaanalysis. This function enables the user to keep the results
of one of these models and to add results from a second
metaanalysis or a sensitivity analysis.
Applications of this function include printing and plotting results of the common effect or random effects metaanalysis and the
trimandfill method (trimfill
),
limit metaanalyis (limitmeta
from R
package metasens),
Copas selection model (copas
from R
package metasens),
robust variance metaanalysis model
(robu
from R package robumeta).
The first argument must be an object created by a metaanalysis
function, e.g., metagen
or metabin
. It
is also possible to provide an object created with
limitmeta
or
copas
. In this case, arguments meta2
and pooled2
will be ignored.
The second metaanalysis could be an object created by a
metaanalysis function or with trimfill
,
limitmeta
, copas
,
or robu
.
The created metaanalysis object only contains the study results
from the first metaanalysis which are shown in printouts and
forest plots. This only makes a difference for metaanalysis
methods where individual study results differ, e.g.,
MantelHaenszel and Peto method for binary outcomes (see
metabin
).
R function metabind
can be used to print and plot the
results of more than two metaanalyses, however, without showing
individual study results.
An object of class "meta"
and "metamerge"
with
corresponding generic functions (see metaobject
).
The following list elements have a different meaning:
TE, seTE, studlab 
Treatment estimate, standard error, and study labels (first metaanalyis). 
lower, upper 
Lower and upper confidence interval limits for individual studies (first metaanalysis). 
statistic, pval 
Statistic and pvalue for test of treatment effect for individual studies (first metaanalysis. 
w.common 
Weights of first common effect metaanalysis. 
w.random 
Weights of first random effects metaanalysis. 
k 
Number of studies combined in first metaanalysis. 
Furthermore, metaanalysis results of common effect or random
effects model are taken from first metaanalysis if only random
effects or common effects models are selected from both
metaanalyses (arguments pooled1
and pooled2
).
Guido Schwarzer sc@imbi.unifreiburg.de
metagen
, metabind
data(Fleiss1993cont) # m1 < metacont(n.psyc, mean.psyc, sd.psyc, n.cont, mean.cont, sd.cont, data = Fleiss1993cont, sm = "MD", text.random = "Random effects model (REML)", text.w.random = "DL") # # Use DerSimonianLaird estimator of tau2 # m2 < update(m1, method.tau = "DL", common = FALSE, text.random = "Random effects model (DL)", text.w.random = "DL") # # Merge results of the two metaanalyses # m12 < metamerge(m1, m2) m12 forest(m12, rightcols = c("effect", "ci", "w.common")) # Show in addition the results for the PauleMandel estimate of # betweenstudy variance # m3 < update(m1, method.tau = "PM", text.random = "Random effects moded (PM)", text.w.random = "PM") # m123 < metamerge(m12, m3, pooled2 = "random") m123 data(Fleiss1993bin) # # MantelHaenszel method # m4 < metabin(d.asp, n.asp, d.plac, n.plac, data = Fleiss1993bin, studlab = paste(study, year), sm = "OR", random = FALSE, text.common = "MH method", text.w.common = "MH") # # Peto method # m5 < update(m4, method = "Peto", text.common = "Peto method", text.w.common = "Peto") # # Merge results (show individual results for MH method) # m45 < metamerge(m4, m5) summary(m45) forest(m45, digits = 4) # # Merge results (show individual results for Peto method) # m54 < metamerge(m5, m4) summary(m54) forest(m54)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.