[R-meta] error for meta-regression with continuous moderator

[Guido Schwarzer]

David,

the command metaprop(..., subgroup = moderator) does not conduct a 
mixed-effects meta-regression with a continuous moderator, but a 
subgroup analysis. In your data example this means that you try to 
estimate 8 subgroup random effects from 14 studies (with 6 subgroups 
having just one study). Accordingly, it is not astonishing that you get 
a warning message (which is not an error message).

I do not get any warnings conducting a meta-regression for your data 
(using meta, version 2.0).

event<-c(28,50,59,29,32,100,36,29,9,50,54,43,49,24)
total<-c(55,60,65,43,41,107,40,34,10,58,54,50,60,28)
moderator<-c(5,5,1,11,9,9,5,9,12,2,3,7,5,9)
data<-data.frame(event,total,moderator)


## Meta-analysis (without covariate)
##
meta <- metaprop(event = event, n = total, data = data, fixed = FALSE)
meta


## Meta-regression using continuous covariate
##
metareg(meta, ~ moderator)


## Subgroup analysis assuming common tau2 in subgroups (categorical 
covariate)
##
print(update(meta, subgroup = moderator, tau.common = TRUE),
       backtransf = FALSE)
metareg(meta, ~ as.factor(moderator) - 1)

Best, Guido