Skip to content

[R-meta] modifying the default forest plot

5 messages · MK, Ken Beath, Gerta Ruecker +2 more

Original
MK
# 
Hi everyone!

I am new to meta-analysis methodology and I am using R software to do my
quantitative synthesis. I am using random effects model to analyse my
individual studies through rma function in metafor package.

By using 'forest(mymodelname)', I have a basic forest plot where my
individual studies are shown as Study 1, Study 2, and so on. How could I
get the author and year instead? It would be better if I can get them
arranged according to the observed effect size - is that possible? Also, I
would like to change the title of my X axis to 'Estimated Effect Size'
rather than the default title 'Observed Outcome'. Lastly, would it be
possible to specify the exact weight assigned to each study in the forest
plot?
Many thanks in advance for your help.


MK
Ken Beath
# 
In the calls there is an slab parameter, use that to label your studies.

To get ordered studies sort them first, before calling the metafor routine.

metafor takes the standard plot parameters so just adding them at the end should work. Look up plot in the help.

I?m not certain what to do for the last as I?m not quite certain what you are asking for.

Ken
Gerta Ruecker
# 
Hi,

What about using the R package meta that has the most flexible forest 
function? You can specify all sorts of things, such as ordering, labels, 
title, x-axis label, marker type, fonts, colours, whether the weights 
should be plotted, and so on. See 
https://cran.r-project.org/web/packages/meta/index.html .

Best,

Gerta


-- Dr. rer. nat. Gerta R?cker, Dipl.-Math. Medical Faculty and Medical 
Center - University of Freiburg Institute for Medical Biometry and 
Statistics Stefan-Meier-Strasse 26, D-79104 Freiburg, Germany Phone +49 
(0)761 2036673 Fax +49 (0)761 2036680 Mail ruecker at imbi.uni-freiburg.de 
Web www.imbi.uni-freiburg.de/biom/

Am 05.12.2017 um 07:42 schrieb Ken Beath:

  
  


  


      

    

    

      
      

        


  

        

      Guido Schwarzer
    

    

      
      #
    

  


  

      
Am 05.12.17 um 09:58 schrieb Gerta Ruecker:

            

      
      
      
    

        
Just to give some examples. The forest plots in the attached PDF-file 
were generated using the following R commands:

library(meta)

set.seed(666)
##
mean <- 1:5
se.mean <- mean / 10
author <- sample(LETTERS, 5)
year <- 2005 + sample(-10:10, 5)

m1 <- metagen(mean, se.mean, sm = "MD",
 ????????????? studlab = paste(author, year))


pdf("forest18.pdf", width = 9.75, height = 4)
##
forest(m1)
forest(m1, sortvar = -TE)
forest(m1, sortvar = year)
forest(m1, xlab = "Estimated Effect Size")
forest(m1, test.overall = TRUE, colgap.forest.left = "2cm")
forest(m1, layout = "JAMA")
forest(m1, layout = "RevMan5")
##
dev.off()


You can use the arguments 'leftcols' and 'rightcols' to add / remove 
columns from the left and right side of the forest plot.

Best wishes,
Guido

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Viechtbauer Wolfgang (STAT)
    

    

      
      #
    

  


  

      
Ken already gave some pointers.

1) How could I get the author and year instead? 

Use the 'slab' argument. For example:

slab=paste(dat$author, dat$year, sep=", "))

(assuming that there are 'author' and 'year' variables in the data frame 'dat').

2) It would be better if I can get them arranged according to the observed effect size - is that possible? 

Use argument: order="obs"

3) Also, I would like to change the title of my X axis to 'Estimated Effect Size'
rather than the default title 'Observed Outcome'. 

Use argument: xlab="Estimated Effect Size"

4) Lastly, would it be possible to specify the exact weight assigned to each study in the forest plot?

Use argument: showweights=TRUE

Best,
Wolfgang

-----Original Message-----
From: R-sig-meta-analysis [mailto:r-sig-meta-analysis-bounces at r-project.org] On Behalf Of Mukht Kaur
Sent: Tuesday, 05 December, 2017 6:30
To: r-sig-meta-analysis at r-project.org
Subject: [R-meta] modifying the default forest plot

Hi everyone!

I am new to meta-analysis methodology and I am using R software to do my
quantitative synthesis. I am using random effects model to analyse my
individual studies through rma function in metafor package.

By using 'forest(mymodelname)', I have a basic forest plot where my
individual studies are shown as Study 1, Study 2, and so on. How could I
get the author and year instead? It would be better if I can get them
arranged according to the observed effect size - is that possible? Also, I
would like to change the title of my X axis to 'Estimated Effect Size'
rather than the default title 'Observed Outcome'. Lastly, would it be
possible to specify the exact weight assigned to each study in the forest
plot?
Many thanks in advance for your help.

MK