I would like to know if anyone knows how to make a forest plot in R where the ORs, the N and the percentage of the control and experimental group are included ... and if possible that this is segmented into subgroups.
Thank you very much in advance! Much success to all with your projects.
Here's some code using ggplot. Note that you have to specify xmin and xmax for ggplot to identify the confidence interval range. geom_pointrange() is the key to the forestplot aesthetic.
ggplot(data, aes(x, y, xmin = conf_low, xmax = conf_hig)) +
geom_pointrange()
edit: sorry, I’ve just seen your request for numbers, etc. you may want to check out the metaviz package, which I think allows you to suppress the summary figure (so it wouldn’t be a meta-analysis). metaviz can also give you subgroups plus tabular information.
Hello Edgar,
Since you apparently don't want an overall estimate (noting that your query indicates this is not meta-analysis), here's a link that describes what essentially is a forest plot method for multiple outcomes, that would be pertinent for odds ratios (if scale centered at 1) or log-odds ratios (if scale centered at zero):
https://finalfit.org/reference/coefficient_plot.html
Good luck with your work.
What I found so far as the best way to do a forest plot without being a meta analisis, is the following.
In RStudio you upload the following packages: ggplot2 and ggpubr. Add the variable, the OR/AOR/RR/whatever, the lower limit and the upper limit all in separate columns in an excel spreadsheet. Upload the excel to R.
Then you use the following:
fp
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis