[R-meta] Meta package - add PI to each subgroup in a forest plot

[Dr. Guido Schwarzer]

> Am 08.11.2022 um 22:18 schrieb Jorge Teixeira <jorgemmtteixeira using gmail.com>:
> 
> Dear all,
> 
> Is it possible to add a Prediction interval to each subgroup in a forest
> plot?

Yes, since version 5.2-0 of meta, using argument ‚prediction.subgroup‘, e.g.,

m <- metagen(1:10, 1:10 / 10, subgroup = rep(LETTERS[1:4], 1:4))
##
forest(m, addrows.below.overall = 2)
forest(m, addrows.below.overall = 2, prediction = TRUE)
forest(m, addrows.below.overall = 2, prediction.subgroup = TRUE)
forest(m, addrows.below.overall = 2, prediction.subgroup = k.w > 3)
forest(m, addrows.below.overall = 2, prediction.subgroup = TRUE,
overall = FALSE)
##
forest(update(m, prediction.subgroup = TRUE, overall = FALSE))

Best wishes,
Guido

P.S. You are still using the deprecated arguments ‚comb.random‘ and ‚byvar‘ which have been replaced by ‚random‘ and ’subgroup’ in version 5.0-0 of meta, see output of news(package = "meta“) for more details on changes in meta.