I want to assess and compare the validity of two prediction models in predicting mortality with my patients. According to the other models, I have I have calculated risk of mortality (p) and logit. But I don't know how to assess calibration with a Cox model (using Hosmer-Lemeshow goodness-of-fit test) analysis using a logistic regression model to obtain two derivatives-the slope and the intercept. I asked a lot of my friends for help, but they have no idea.
just like the follow pictures
Asked
Active
Viewed 83 times
1
-
Can you say more about your models? Are you comparing a Cox model to a logistic regression model? – gung - Reinstate Monica Aug 03 '15 at 20:35
-
What software are you using? – Jake Aug 03 '15 at 20:46
-
thank you for all of you.I just want to validation two logistic regression models.And I learned from other authors that they use a Cox model to assess the calibration. – lei chen Aug 03 '15 at 20:59
-
Hosmer-Lemeshow is considered obsolete : http://stats.stackexchange.com/questions/145112/hosmer-lemeshow-test-with-weighted-data/145116#145116 – kjetil b halvorsen Apr 13 '17 at 09:27