Dear All
I have a survival data that I divided into training and test dataset.
I ran a cox regression model on the training dataset and evaluated the concordance index on the test dataset. The concordance is 0.56.
I did the same using Random Survival Forests and the concordance index is 0.66
My boss is telling me that I cannot say that RSF is better than the cox model just because the numbers looks better.
He said I need to compare between the two models statistically.
I wonder if any of you guys aware of any test that can do that?
If there is not any, I wonder if improving the concordance index from 0.56 to 0.66 is considered significant? Is there a reference for that?