[R-meta] Peto random effect?

[Guido Schwarzer]

Am 06.10.21 um 12:39 schrieb Philippe Tadger:

> Dear r-sig-meta community,
>
> In Revman, doesn't seems to be options to do a Peto with random effect,
> But I see that is available in meta and metafor.
> [...]
> Am I missing something here? Does any one know the reason why is not
> available in Revman?

I cannot answer the question concerning RevMan.

The Peto odds ratio (POR) for an individual study is constructed under 
the null hypothesis of no treatment difference, i.e., a true odds ratio 
of 1. This implies that the POR is a biased estimate of the OR for large 
treatment effects, i.e., if the true odds ratio is far away from 1 
(Brockhaus et al., 2014).

One may argue that the Peto method, i.e., a meta-analysis using the POR 
as effect measure, is inherently a fixed effect / common effect method 
due to its construction under the null hypothesis of no treatment 
difference. However, in my view, the POR is an effect measure and I do 
not see why it should not be used in a random effects meta-analysis.

Best, Guido

1.
Brockhaus AC, Bender R, Skipka G. The Peto odds ratio viewed as a new 
effect measure. Statistics in Medicine. 2014;33(28):4861–74.