[R-meta] Meta-analysis on BD Model (Budding and Depth of Invasion in Tumor)
Awais Wahab
@||gner@79 @end|ng |rom gm@||@com
Wed Mar 13 13:45:35 CET 2019
Dear Sir/Madam!
I hope You are doing good.
My study is on meta analysis in R, with HR values for each study and CIs.
It is kind request if someone can please answer number wise for better
understanding.
1. Should for univariate meta-analysis, both OS and DFS need to be only one
each for a forest plot Or can we have for example, two DFS and one OS or
vice versa ? (I got above question number 1. earlier as follows,
*"In univariate meta-analysis, each study can only contribute a single *
*hazard ratio, i.e., one univariate meta-analysis for DFS and another *
*univariate meta-analysis for OS.*
*However, it is possible to conduct a multivariate meta-analysis (e.g., *
*using R package mvmeta) taking a single DFS and OS estimate from each *
*study into account (for details, see Riley et al., 2007)."*
2. And similarly like question 1 and above text, does same thing goes for
this multivariate meta-analysis forest plot that we can take. for example,
two DFS and one OS HRs or vice versa ?
3. My one study has both HR with univariate DFS and univariate OS and
second study has univariate DFS only ,so is it better to include two
univariate DFS HR values for forest plot. Or should I report both scenarios
in two forest plot separately?
4. Please let me know whether only two studies HR are enough to perform
meta analysis with forest plot ? I am planning to present two forest plots
for two studies each, like 2 DFS multivariate HR for one forest plot and 2
DFS univariate HR for one forest plot.
5. Is it possible to combine univariate HR of one study with
multivariate/s HR of another study for DFS or DSS etc.?
6. Is there any method to include separate studies or articles of
parameters of tumour budding and depth of invasion separately for HRs and
later merge their results for meta-analysis of combined (budding and depth
of invasion) BD model ?
best regards,
Awais
[[alternative HTML version deleted]]
More information about the R-sig-meta-analysis
mailing list