[R-meta] Fwd:

[Michael Dewey]

If you have the effect sizes (the HR) then what scientific question are 
you hoping to answer by also analysing the p-values? Normally that is a 
last resort when you do not have effect sizes.

Michael

On 06/10/2021 10:03, Im wrote:
> 
> Dear All!
> Thanks a lot everyone for helping me to conduct the meta analyses and 
> making me reach to results stage of the manuscript.
> 
> However, I still have some crucial questions regarding interpretation of 
> meta-analysis with HR values, alongwith meta-analysis of their 
> respective P-values.
> *
> *
> 
> *Data Collection:*
> *
> *
> Mainly, I have gathered *2-4 P values each of Hazard Ratios (HR values 
> given with CIs) from studies on* *specific* *prognostic markers 
> reporting* *DFS, OS and DSS*.
> 
> 
> *_R Studio Codes for meta-analysis of HR values:_*
> *
> *
> *library(meta)*
> *
> settings.meta(digits = 2, layout = "RevMan5")
> *
> *
> *
> *HR <- c(1.34,1.95) */(Hypothetical values)/
> *
> lower.HR <- c(1.10,0.75) *
> *
> upper.HR <- c(2.34,2.25) *
> *
> study<- c("ABC et al. 2018","EFG et al. 2020")*
> *
> m2<-metagen(log(HR),lower=log(lower.HR),upper=log(upper.HR),studlab=study,sm="HR")*
> *
> m2*
> *
> forest(m2)*
> *
> grid::grid.text("Meta-analysis of Prognostic marker for Multivariate DFS",
>                  0.5, 0.94, gp = grid::gpar(cex = 1.5)) *
> 
> 
> *_R Studio Codes for meta-analysis of P values:_*
> *
> *
> 
> *install.packages("BiocManager")*
> 
> *BiocManager::install("multtest")*
> 
> *install.packages("metap")*
> 
> **
> 
> *library(metap)*
> 
> *pvals <- c(0.08,0.3) */(Hypothetical values)/
> 
> *//*
> 
> *result <- sumlog(pvals)*
> 
> **
> 
> *result*
> 
> *
> *
> 
> *_Results of Meta-analysis with P values:_*
> *
> *
> 
> 1. chisq = 16.66974  with df =  4  p =  0.002240369
> 
> 
> 2. chisq =  35.80921 with df =  8  p = 1.902998e-05   (**Do we report 
> this result as 0.000019 in the manuscript ?)
> 
> 
> 
> *_Questions:_*
> 
> *
> *
> 
> 1.Please view attached first image of forest plot which I have got from 
> "Meta" Package in RStudio. The P value in this forest plot is for 
> heterogeneity, but the P value for overall effect size is not there as 
> presented in the second attached image as: *(Test or overall effect 
> z=0.35, /p/=0.7)*. Should I, and if "yes" then how to report this P 
> value in my forest plot ?
> 
> And is this P value similar to the P value which I am getting from 
> meta-analysis of P values with "metap" package ? which P value to be 
> reported ideally and usually ?
> 
> 
> 2. Regarding *"metap*" package my cumulative P values are too low, so 
> how to interpret these results. Is there no image OR forest plot for 
> them to report ? Does the significance level (0.005) changes for 
> cumulative P values ?
> 
> 
> 3. Can I report these cumulative P value meta-analysis results along 
> with regular meta-analysis of cumulative HR values (forest plots) ?
> 
> 
> I will be much grateful for your much valuable expert advice to finalize 
> my results.
> best regards,
> Imran
> 
> <http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=emailclient> 
> 	Virus-free. www.avg.com 
> <http://www.avg.com/email-signature?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=emailclient> 
> 
> 
> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
> 
> _______________________________________________
> R-sig-meta-analysis mailing list
> R-sig-meta-analysis using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis
> 

-- 
Michael
http://www.dewey.myzen.co.uk/home.html