[R-meta] combining Means + SD at post, withD + 95% CI

[Dr. Gerta Rücker]

Dear Jorge,

First of all, your formula for the variance of the XING study does not seems correct to me. I would expect for the sampling variance:

SE^2 = (CI.length / 1.96 /2)^2 = (8 / 1.96 / 2)^2 = 4.164931

(It is correct that you don't need the sample size.)

This gives with metafor:
estimate      se    zval    pval    ci.lb   ci.ub    
  2.0751  1.0832  1.9156  0.0554  -0.0481  4.1982  .

Now how it works with meta:

# Your data
dat <- structure(list(study = c("AA", "BB", "CC", "EE", "DDD"), 
                      en = c(9, 41, 29, 8, 13), 
                      em = c(32, 27.5, 28.7, 22.8, 30.5), 
                      esd = c(1.9, 3.8, 5.2,  3.8, 4.9), 
                      cn = c(8, 26, 28, 10, 14), 
                      cm = c(30.1, 24.9, 26.9, 24.7, 30.2), 
                      csd = c(2.4, 3.6, 3, 4.3, 5.1)), 
                 row.names = c(NA,-5L), class = c("tbl_df", "tbl", "data.frame"))
# Meta-analysis of continuous data with metacont()
m0 <- metacont(n.e = en, mean.e = em, sd.e = esd, n.c = cn, mean.c = cm, sd.c = csd, 
               data = dat)
# Use treatment effects (TE) and their standard errors (seTE) from m0, add the CI of study XING and conduct generic inverse variance meta-analysis
# metagen() will do the transformation from CI to TE and seTE
m1 <- metagen(TE = c(m0$TE, NA), seTE = c(m0$seTE, NA),
              lower = c(NA, NA, NA, NA, NA, 4),
              upper = c(NA, NA, NA, NA, NA, 12),
              common = FALSE, method.tau = "REML")
m1
forest(m1)

Best,
Gerta






UNIVERSITÄTSKLINIKUM FREIBURG
Institute for Medical Biometry and Statistics

Dr. Gerta Rücker
Guest Scientist

Stefan-Meier-Straße 26 · 79104 Freiburg
gerta.ruecker using uniklinik-freiburg.de

https://www.uniklinik-freiburg.de/imbi-en/employees.html?imbiuser=ruecker


-----Ursprüngliche Nachricht-----
Von: R-sig-meta-analysis <r-sig-meta-analysis-bounces using r-project.org> Im Auftrag von Jorge Teixeira via R-sig-meta-analysis
Gesendet: Montag, 27. März 2023 16:25
An: Michael Dewey <lists using dewey.myzen.co.uk>
Cc: Jorge Teixeira <jorgemmtteixeira using gmail.com>; R Special Interest Group for Meta-Analysis <r-sig-meta-analysis using r-project.org>
Betreff: Re: [R-meta] combining Means + SD at post, withD + 95% CI

 Dear Michael, thank you - that is indeed more elegant.

Out of curiosity, is it possible to do this in “meta” package too?

Michael Dewey <lists using dewey.myzen.co.uk> escreveu no dia segunda, 27/03/2023
à(s) 13:32:

> Dear Jorge
>
> I think the conv.wald() function in metafor will calculate the variance
> for you. You do need an up-to-date version of metafor.
>
> Michael
>
> On 26/03/2023 21:28, Jorge Teixeira via R-sig-meta-analysis wrote:
> >   Hi all.
> >
> > I did this to combine studies that provided M+SD, with others for which I
> > had only MD + 95% CI or p-value.
> >
> >
> > *Questions: *
> >
> > i)   Is the Vi well calculated below, as the code was given by GPT?  n
> per
> > group is irrelevant here, right?
> >
> > ii)  Is there a more elegant way to do this?
> >
> > iii) Can this method backfire in some way or is there something in
> > particular that I should have in mind?
> >
> >
> > Thanks! Code below:
> >
> > library(metafor)
> >
> > #
> > dat <- structure(list(study = c("AA", "BB", "CC",
> >                                  "EE", "DDD"), en = c(9, 41, 29, 8, 13
> >                                  ), em = c(32, 27.5, 28.7, 22.8, 30.5),
> esd
> > = c(1.9, 3.8, 5.2,
> >
> >     3.8, 4.9), cn = c(8, 26, 28, 10, 14), cm = c(30.1, 24.9, 26.9,
> >
> >                                                  24.7, 30.2), csd =
> c(2.4,
> > 3.6, 3, 4.3, 5.1)), row.names = c(NA,
> >
> >
> >                                  -5L), class = c("tbl_df", "tbl",
> > "data.frame"))
> >
> > #
> > dat <- escalc(measure="MD", m1i=em, sd1i=esd, n1i=en, m2i=cm, sd2i=csd,
> > n2i=cn, data=dat)
> >
> >
> > # Now lets had study "XING", which has a MD of 8 (95%CI: 4, 12), n=60;
> > p=0.001
> >
> > sampling variance = (CI / 1.96) ^ 2 / (2 * sample size)
> >
> > (8 / 1.96) ^ 2 / (2 * 6) = 1.388
> >
> >
> > # Create a new row for study XING
> > new_row <- data.frame(study = "XING", en = NA, em = NA, esd = NA, cn =
> NA,
> > cm = NA, csd = NA, yi = 8, vi = 1.388)
> >
> > # Add the new row to dat
> > dat <- rbind(dat, new_row)
> >
> > #
> > res <- rma(yi, vi,  data=dat)
> > res
> >
> >       [[alternative HTML version deleted]]
> >
> > _______________________________________________
> > R-sig-meta-analysis mailing list @ R-sig-meta-analysis using r-project.org
> > To manage your subscription to this mailing list, go to:
> > https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis
> >
>
> --
> Michael
> http://www.dewey.myzen.co.uk/home.html
>

	[[alternative HTML version deleted]]

_______________________________________________
R-sig-meta-analysis mailing list @ R-sig-meta-analysis using r-project.org
To manage your subscription to this mailing list, go to:
https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis