CODEWITHSUNDEEP
rr-forestplot
Nicolas
Nicolas

Reputation: 444

Is there a way in R to present a custom summary of effect sizes in a meta or forest object?

I need to present a metanalysis of vaccine effectiveness studies. In essence, I need the summary statistics presented to be as vaccine effectiveness or "VE" which is "1-RR" instead of "RR", or alternatively, "1-OR" instead of "OR". Currently, the meta objects allow for these standard effect sizes size as RR and OR, but is there a way of presenting a custom output?

The Metafor package has a large list of transformations (see "transf" option), none of which allow custom inputs, and the the "Meta" package also has no custom inputs.

Is there another program, or method to be aware of?

It might look like the following (using data from the metafor package):

### load metafor package
library(metafor)

### load BCG vaccine dataset
data(dat.bcg)
### Example of what it might look like with a custom summary estimate
dat <- escalc(measure="1-RR", ai=tpos, bi=tneg, ci=cpos, di=cneg, data=dat.bcg)

Upvotes: 0

Views: 122

Answers (1)

Wolfgang
Wolfgang

Reputation: 3395

In the predict() function of metafor, you can specify custom transformations. So you conduct the meta-analysis with the (log-transformed) risk ratios and then transform as you see fit. For example:

library(metafor)

# compute log risk ratios and corresponding sampling variances
dat <- escalc(measure="RR", ai=tpos, bi=tneg, ci=cpos, di=cneg, data=dat.bcg)

# fit random-effects model
res <- rma(yi, vi, data=dat)
res

# estimated pooled risk ratio (with 95% CI/PI)
predict(res, transf=exp)

# estimated pooled 1-RR (with 95% CI/PI)
predict(res, transf=function(x) 1-exp(x))

Upvotes: 1

Related Questions