0

I have some difficulties using metafor from R.

I am doing a prevalence meta-analysis and I used arcsine transformation. Then for the predicted values I need to use the argument transf=iarcsin to back transform to a prevalence rate. For the forest plot, I want to have also the prevalence rate according to the calculated weight corresponding to each study, but I only get the arcsine transformed value, no matter whether I use the transf argument or not, I get the same result (yi instead of prevalence rates in the forest plot).

This is the code I am using:

forest(alex1.ma, addcred = TRUE, header = TRUE, transf = "iarcsine", slab = authors, showweights = TRUE, ilab = alex1_df$css, ilab.xpos = c(-0.5))

Thanks

TBG
  • 1
  • 2
    It's easier to help you if you include a simple [reproducible example](https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example) with sample input and desired output that can be used to test and verify possible solutions. – MrFlick Sep 03 '20 at 20:54
  • You need to use `transf = transf.iarcsin`. – Wolfgang Sep 04 '20 at 08:17
  • Dear Wolfgang, I had tried `transf = transf.iarcsin` but it is the same output than using transf =iarcsin or not using them – TBG Sep 07 '20 at 12:17
  • You will have to post a fully reproducible example. For example, `dat <- escalc(measure="PAS", xi=c(10,20), ni=c(50,50)); res <- rma(yi, vi, data=dat); forest(res, transf=transf.iarcsin)` works just fine. – Wolfgang Sep 20 '20 at 09:42
  • I hope this is useful. `alex1.transf <- escalc(measure="PAS",xi=css_n,ni=css_N,data=alex1_df, add=0)` ; `alex1.ma <- rma(yi,vi,data = alex1.transf, method="PM",weighted=TRUE)` ; `forest(alex1.ma, addcred = TRUE, header = TRUE, transf = "iarcsine", slab = names, ilab = alex1_df$css, ilab.xpos = c(-0.25), cex=.7)` – TBG Sep 28 '20 at 13:20
  • `pes <- predict(alex1.ma, transf=transf.iarcsin)` gives me a pred = 0.2814, but the forest plot as I posted shows 0.58 and it is the same if I skip `transf = "iarcsine"` in the coding. I am performing arcsine transformation (not double-arcsine). As I understand, the problem is that the retro-transformation is not done when ploting – TBG Sep 28 '20 at 13:26

0 Answers0