[R-meta] Zero seTE in network MA

[Guido Schwarzer]

Dear Carla,

I think this is a misunderstanding.

R function netmeta() works perfectly with zero standard errors (by 
excluding pairwise comparisons with zero or missing standard errors from 
the network meta-analysis). This is the same behaviour as in "basic" 
meta-analysis of pairwise comparisons (see metagen() in R package 
*meta*). I attached a corresponding fictitious example for netmeta() to 
this email.

The new checks in *netmeta*, version 0.9-6, you are referring to, 
related for /treatment arm variances in multi-arm studies/ which are 
calculated internally. Sometimes, users enter standard errors for 
pairwise comparisons in multi-arm studies that simply do not "add up". 
Accordingly, treatment arm variances can be negative.

Maybe somebody else can also comment on your idea to use a (very) small 
standard error for a pairwise comparison with zero standard error. In my 
view, the problem with such an approach is that this study gets a very 
large weight in the (network) meta-analysis.

Best wishes,
Guido

-- 
Dr. Guido Schwarzer (sc at imbi.uni-freiburg.de)
Institute for Medical Biometry and Statistics
Stefan-Meier-Str. 26, D-79104 Freiburg | Phone: +49 (0)761 203 6668
http://www.imbi.uni-freiburg.de        | Fax:   +49 (0)761 203 6680

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://stat.ethz.ch/pipermail/r-sig-meta-analysis/attachments/20170920/bb9adf15/attachment.html>
-------------- next part --------------
library(netmeta)

p1 <- pairwise(treat = c(LETTERS[1:3], LETTERS[2:3]),
               n = rep(10, 5), mean = 1:5, sd = 0:4,
               studlab = c(rep("Study 1", 3), rep("Study 2", 2)))

netmeta(p1, ref = "C")

p1$seTE[is.na(p1$seTE)] <- 0

netmeta(p1, ref = "C")
-------------- next part --------------

R version 3.4.1 (2017-06-30) -- "Single Candle"
Copyright (C) 2017 The R Foundation for Statistical Computing
Platform: x86_64-apple-darwin16.7.0 (64-bit)

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

[Previously saved workspace restored]

> library(netmeta)
Loading required package: meta
Loading 'meta' package (version 4.8-4).
Type 'help(meta)' for a brief overview.
Loading 'netmeta' package (version 0.9-6).
Type 'help(netmeta-package)' for a brief overview.
> 
> p1 <- pairwise(treat = c(LETTERS[1:3], LETTERS[2:3]),
+                n = rep(10, 5), mean = 1:5, sd = 0:4,
+                studlab = c(rep("Study 1", 3), rep("Study 2", 2)))
Comparisons will not be considered in network meta-analysis:
 TE seTE studlab treat1 treat2 n1 mean1 sd1 n2 mean2 sd2
 -1   NA Study 1      A      B 10     1   0 10     2   1
 -2   NA Study 1      A      C 10     1   0 10     3   2
Warning messages:
1: In metacont(dat$n1, dat$mean1, dat$sd1, dat$n2, dat$mean2, dat$sd2,  :
  Studies with non-positive values for sd.e or sd.c get no weight in meta-analysis.
2: In metacont(dat$n1, dat$mean1, dat$sd1, dat$n2, dat$mean2, dat$sd2,  :
  Studies with non-positive values for sd.e or sd.c get no weight in meta-analysis.
3: Comparisons with missing TE / seTE or zero seTE will not be considered in network meta-analysis. 
> 
> netmeta(p1, ref = "C")
Comparisons not considered in network meta-analysis:
 studlab treat1 treat2 TE seTE
 Study 1      A      B -1   NA
 Study 1      A      C -2   NA
Original data:

        treat1 treat2 TE   seTE
Study 1      B      C -1 0.7071
Study 2      B      C -1 1.5811

Number of treatment arms (by study):
        narms
Study 1     2
Study 2     2

Results (fixed effect model):

        treat1 treat2 MD             95%-CI Q leverage
Study 1      B      C -1  [-2.2652; 0.2652] 0     0.83
Study 2      B      C -1  [-2.2652; 0.2652] 0     0.17

Number of studies: k = 2
Number of treatments: n = 2
Number of pairwise comparisons: m = 2
Number of designs: d = 1

Fixed effect model

Treatment estimate (sm = 'MD', comparison: 'B' vs 'C'):
  MD             95%-CI
B -1  [-2.2652; 0.2652]
C  .                  .

Quantifying heterogeneity / inconsistency:
tau^2 = 0; I^2 = 0%

Test of heterogeneity / inconsistency:
 Q d.f.  p-value
 0    1   1.0000
Warning messages:
1: Comparisons with missing TE / seTE or zero seTE not considered in network meta-analysis. 
2: Only a single design in network meta-analysis. 
> 
> p1$seTE[is.na(p1$seTE)] <- 0
> 
> netmeta(p1, ref = "C")
Comparisons not considered in network meta-analysis:
 studlab treat1 treat2 TE seTE
 Study 1      A      B -1    0
 Study 1      A      C -2    0
Original data:

        treat1 treat2 TE   seTE
Study 1      B      C -1 0.7071
Study 2      B      C -1 1.5811

Number of treatment arms (by study):
        narms
Study 1     2
Study 2     2

Results (fixed effect model):

        treat1 treat2 MD             95%-CI Q leverage
Study 1      B      C -1  [-2.2652; 0.2652] 0     0.83
Study 2      B      C -1  [-2.2652; 0.2652] 0     0.17

Number of studies: k = 2
Number of treatments: n = 2
Number of pairwise comparisons: m = 2
Number of designs: d = 1

Fixed effect model

Treatment estimate (sm = 'MD', comparison: 'B' vs 'C'):
  MD             95%-CI
B -1  [-2.2652; 0.2652]
C  .                  .

Quantifying heterogeneity / inconsistency:
tau^2 = 0; I^2 = 0%

Test of heterogeneity / inconsistency:
 Q d.f.  p-value
 0    1   1.0000
Warning messages:
1: Comparisons with missing TE / seTE or zero seTE not considered in network meta-analysis. 
2: Only a single design in network meta-analysis. 
>