[R-meta] Facing some issues on Meta - regression

[Dr. Gerta Rücker]

Dear Sancho,

It is correct that the back-transformation should be the inverse of the logit, which is plogis().

Thank you for sending data and code. However, the code works only after some changes:
1. data set called hiper does not exist, the data you provided is called OpioidMisuse
2. column events does not exist, the correct spelling is event
3. column id_article does not exist, has to replaced with author
4. column population does not exist, has to be replaced with n
5. Year_publi is not part of the data set and is defined in your code only after it is needed

After fixing all this, I could run your code. 
In the spirit of meta, transformatuon of the y-axis is meant to work automatically (or after using the argument backtransf, as Michael wrote). However, as I already wrote, this does not work with metaprop(), for reasons I don't know. The call would be

bubble(m.prop.reg, studlab = TRUE, backtransf = TRUE)

Thus, as said before, this is a question to @Guido who maintains the meta package.

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

Von: Sancho Xavier via R-sig-meta-analysis <r-sig-meta-analysis using r-project.org> 
Gesendet: Montag, 12. August 2024 15:19
An: Michael Dewey <lists using dewey.myzen.co.uk>; R Special Interest Group for Meta-Analysis <r-sig-meta-analysis using r-project.org>
Cc: Sancho Xavier <sanchoxavierxavier using gmail.com>
Betreff: [SPAM](FN) Re: [R-meta] Facing some issues on Meta - regression

It seems to me that the specific function sm = "PLOGIT" for calculating proportions using the number of events and the number of observations (populations) in the metaprop function employs a logit transformation.
Therefore, to represent the proportions in the figure (meta-regression) without negative values on the X-axis, it would be necessary to apply the inverse logit transformation (using transf.ilogit). I'm not sure if this is scientifically the correct approach (applying the inverse logit transformation).

You can use the dataset and de codes below:
Dataset called => opioidmisuse
codes => Testing_sharing

Sancho Xavier
Candidate for Postgraduate Programme in Epidemiology in Public Health
MBA in Data Science and Analytics at USP (ongoing)
Linkedin: linkedin.com/in/sancho-xavier-59565a18a
Orcid iD: https://orcid.org/0000-0001-9493-4098
Web of Science (Clarivate): https://www.webofscience.com/wos/author/record/GON-5675-2022
Reference Citation Analysis (RCA): https://referencecitationanalysis.com/00001471
Whatsapp: +258 852255846 
------------------------------------------------------------------------------
________________________________________
De: Michael Dewey <mailto:lists using dewey.myzen.co.uk>
Enviado: segunda-feira, 12 de agosto de 2024 07:04
Para: R Special Interest Group for Meta-Analysis <mailto:r-sig-meta-analysis using r-project.org>
Cc: Sancho Xavier <mailto:sanchoxavierxavier using gmail.com>
Assunto: Re: [R-meta] Facing some issues on Meta - regression 
 
It would be easier to help you if you (a) showed your data using dput(), 
(b) pasted your actual code rather than showing a screenshot which means 
having to type it all in again.

Having said that I believe that the plot functions in meta have an 
argument to backtransform the estimates. From the help for bubble

Backtransf

A logical indicating whether results for relative summary measures 
(argument sm equal to "OR", "RR", "HR", or "IRR") should be back 
transformed. If backtransf=TRUE, results for sm="OR" are printed as odds 
ratios rather than log odds ratios, for example.

Michael

On 11/08/2024 20:28, Sancho Xavier via R-sig-meta-analysis wrote:
> Hi everyone, I hope you're all doing well.
> I need your help with something. I'm applying a meta-regression on a 
> dataset where I initially used the |metaprop| function to obtain the 
> pooled proportions of hypertension. Now, I'm trying to understand how 
> variables like the year of publication, study type, and publication 
> pathways influence the effect size (proportion) in the meta-analysis.
> Dataset (called "hiper")
> 
> Here are the scripts I used:
> First, I applied |metaprop|:
> 
> Next, I applied the meta-regression:
> 
> After generating the plot (see below), I noticed that without 
> transforming the |TE| and |seTE| values, the proportions on the x-axis 
> appear as negative values. I would like to know if my scripts for the 
> meta-regression are correct and whether it would be necessary to apply a
> logarithmic transformation (|transf.ilogit(m.prop$TE)| and 
> |transf.ilogit(m.prop$seTE)|) to fix this.
> Additionally, I would appreciate your advice on which method is more 
> appropriate: using the transformed values or the non-transformed ones?
> 
>   Thanks in advance for your help!
> Best wishes,
> 
> *Sancho Xavier*
> Candidate for Postgraduate Programme in Epidemiology in Public Health
> MBA in Data Science and Analytics at USP (ongoing)
> *Linkedin:* linkedin.com/in/sancho-xavier-59565a18a
> *Orcid iD:* https://orcid.org/0000-0001-9493-4098
> *Web of Science (Clarivate):* 
> https://www.webofscience.com/wos/author/record/GON-5675-2022 
> <https://www.webofscience.com/wos/author/record/GON-5675-2022>
> *Reference Citation Analysis (RCA): 
> *https://referencecitationanalysis.com/00001471 
> <https://referencecitationanalysis.com/00001471>
> *Whatsapp: *+258 852255846
> ------------------------------------------------------------------------------
> 
> _______________________________________________
> R-sig-meta-analysis mailing list @ mailto: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

-- 
This email has been checked for viruses by Avast antivirus software.
http://www.avast.com