Timeline for Cox regression baseline risk and interpretation of coefficients
Current License: CC BY-SA 4.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 18 at 15:32 | comment | added | EdM |
@Ania6789 then proceed with the full gene*grade interaction model, which addresses the concerns that you raised in the gene:grade model and provides a direct test (based on the interaction coefficient) of whether the association between gene and outcome depends on grade (and vice-versa). Consider modeling continuous gene values flexibly, and perform appropriate multiple-comparison corrections. You might handle missing clinical data with multiple imputation.
|
|
| Jul 18 at 7:52 | comment | added | Ania6789 | Thank you for your answer! It made it much clearer for me. For the post I was trying to simplify my problem, so there is a mistake (I am really sorry for that). So there are only two grades (2 and 3 are the same as A and B, I have already made a correction). Regarding the number of genes I have a small number o preselected genes and I the clinical characteristic is very limited with a lot of missing data.... | |
| Jul 17 at 16:42 | history | answered | EdM | CC BY-SA 4.0 |