Hello,
I was wondering if there was a way to calculate interaction p-value by hand using estimates from a published article that does not report the p-value??
For instance, if the HR (95% CI) between smoking and lung cancer in Men is 12.3 (95% CI: 3.8 to 17.5) and Women is 8.7 (95% CI: 2.9 to 13.1)
Then I can back-calculate the SE from the CI and see if the HRs for Men and Women vary more than is due by chance alone?
(note: the example above is completely made up)
Hello Spencer,
Yes, you could do that, but note that SE estimates should be derived from log values, not point values or CI limit values. This is explained in a previous RGate post:
https://www.researchgate.net/post/How-to-calculate-Standard-Error-provided-Hazard-Ratio-and-Confidence-Interval-Upper-limit-and-lower-limit-at-95
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Mathematics