For dichotomous variables analysis in the meta-analysis; is there any recommendation when to use Odd ratio versus Risk ratio for the analysis?. I came across a recommendation to use RR when all included studies are RCTs and Odd ratio when they are observational studies or mixed ( observational and RCTs, but I am sure if that is correct or not.
I think from Nov 2020 onwards, the only valid approach is the use of the odds ratios for meta-analysis. However meta-analytic results can be transformed to risks for interpretation - post hoc transformations.
https://www.jclinepi.com/article/S0895-4356(20)31171-9/fulltext
If they are all longitudinal studies (RCT or cohort) use RR.
If case referent studies are included, and rare disease assumption holds in all populations included in meta analysis, then use ORs in case referent studies as approximations to RR. ie do meta analysis on RRs.
Article Meta-Analyze Dichotomous Data: Do the Calculations with Log ...
Ali Yasen Y Mohamedahmed
I think from Nov 2020 onwards, the only valid approach is the use of the odds ratios for meta-analysis. However meta-analytic results can be transformed to risks for interpretation - post hoc transformations.
https://www.jclinepi.com/article/S0895-4356(20)31171-9/fulltext
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis