I am working on a meta-analysis (physical inactivity and the risk of dysmenorrhea); So the calculated ORs physical inactivity/physical activity) yet, in some studies the ORs are calculated for physical activity/physical inactivity.. Then, how to invert these ORs and CIs?
Thank you
If some studies report the odds ratio by calculating the odds as #events/#non-events and others calculate the odds as #non-events/#events, you can convert to a common format by taking the inverse of the format you don't want. Converting the confidence interval is done similarly, i.e., take the inverse of the upper and lower limits. However, be aware that when you take the inverse of the upper limit for one format, it will become the lower limit for the other format. Likewise, when you take the inverse of the lower limit for one format, it becomes the upper limit for the other format.
Thank you for the answer. Gordon L Warren
So, for example OR (95%CI) 2 (1-3) will be 0.5 (0.3-1)?? Is that right?
Yes, you are correct mathematically even though it is impossible to have a symmetric confidence interval around an odds ratio.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis