I would like to know if performing surgery or not is a benefit for a patient. For that, I have a database with baseline characteristics of patients, surgery (and time to surgery) or not and outcome (death). I would like to calculate prediction of deaths depending on the characteristics of the patient and the fact that the patient undergoes or not a surgery. It will be an help for the clinical team to take a decision. I am thinking about an extended Cox model with surgery as time-dependent variable and with significant baseline characteristics included in the model. I found a paper dated on 2005: "Predictive survival model with time-dependent prognosis factors: development of computer-aided SAS macro program". Did someone use it? Do you think that it will answer to my question?
... View more