Forest plot in meta package in R – remove weight boxes

August 12, 2023

I have run a meta-analysis on 10 studies and am now plotting the results on a forest plot. However, I have come across an annoying feature, where the study results which have small SE (fit within the weight box) turn to white, and don’t seem to obey the color.study argument. Is there any other way of bypassing this?
I wouldn’t mind losing the weight boxes altogether if that’s what’s necessary as weights are nearly the same.

df<-data.frame(study=c('s1','s2','s3','s4','s5','s6','s7','s8','s9','s10'),
SMD=c(0.016416141,0.299488949,0.024292283,0.110415656 ,0.160754906,-0.018112735,0.027915812,-0.025253501,-0.024464375,0.001146802),
se=c(0.013495131,0.021345836,0.010902911,0.006237354,0.014503043,0.012365650,0.009825055,0.006861558,0.018569984,0.021838104))

## Run meta-analysis
m.gen <- metagen(TE = SMD,
                 seTE = se,
                 studlab = study,
                 data = df,
                 sm = "SMD",
                 fixed = FALSE,
                 random = TRUE,
                 method.tau = "REML",
                 hakn = FALSE,
                 prediction = TRUE,
                 title = "Test")
    
## Plot forest plot
forest.meta(m.gen, 
            sortvar = TE,
            prediction = TRUE, 
            print.tau2 = FALSE,
            smlab="Test",
            print.I2=FALSE,
            print.I2.ci=FALSE,
            col.study='black'
            
)

>Solution :

From ?forest.meta

col.inside – The colour for individual study results and confidence limits if confidence limits are completely within squares.

So the solution is to set col.inside = "black"

forest.meta(m.gen, 
            sortvar = TE,
            prediction = TRUE, 
            print.tau2 = FALSE,
            smlab = "Test",
            print.I2 = FALSE,
            print.I2.ci = FALSE,
            col.study = 'black',
            col.inside = "black"
)