Color by Significance in R’s forestplot

The forestplot package doesn’t easily support advanced coloring of each point, specifically to flag significance levels.

It does however supported custom coloring via a moderately complicated custom function, so here’s an example of how to color by significance.

First, create the following function:

fn <- local({
  i <- 0
  clrs <- ifelse(dat$pv < .05, "darkgreen",
                 -ifelse(dat$pv < .1, "blue", "darkgrey"))

  function(..., clr.line, clr.marker) {
    i <<- i + 1
                   clr.line = clrs[i],
                   clr.marker = clrs[i])

Inside local() so that i isn’t manipulated elsewhere.

The clrs object defined in lines 3-4 returns a specific color based upon the p-value, here stored in dat$pv. For context, “darkgreen” is used for p-values below .05, “blue” for p-values below .10 (but above .05), and “darkgrey” for p-values above .10.

If you want different colors for the points and the lines, define two different vectors similar to clrs, and change lines 9-10 to point to the differing vectors.

Following the creation of fn, you can add it to your forestplot() via:

           fn.ci_norm = fn,
Home | Back to blog

This work is licensed under CC BY-NC 4.0 Creative Commons BY-NC image