Timeline for Meta Analysis of Cox Regression Coefficients?
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 5, 2015 at 12:38 | comment | added | PickledZebra | Ok, yes that's exactly the type of data I have. Thanks so much. | |
| Oct 5, 2015 at 12:38 | vote | accept | PickledZebra | ||
| Oct 5, 2015 at 5:34 | comment | added | Wolfgang | When survival times are censored. Maybe the event is 'death' and not everybody in the sample has died (yet). But maybe this does not apply in your case and the event of interest has occurred for everybody. At any rate, the SE of the coefficient is all you need. It includes all the information about sample size and the number of events. | |
| Oct 4, 2015 at 21:57 | comment | added | PickledZebra | Yes, this is what I have been doing. I am just worried that it might not be rigorously correct. Can you expand on what you mean by your last sentence? Let's say I have 50 samples, that's 50 measurements and 50 survival times from the corresponding patient. When might the number of events differ from the number of samples (except the trivial case where there is missing data)? | |
| Oct 4, 2015 at 17:15 | history | answered | Wolfgang | CC BY-SA 3.0 |