I'm tasked to perform a Cox Proportional Hazards Regression on our company data. After preparing the data, I started testing the proportional hazards assumption which, according to my readings, will be the main factor in knowing whether cox ph is the appropriate procedure or not.
Unfortunately for me, the survival curves keep on crossing (which, from my understanding, means that I have failed to meet the assumption). Could anyone please tell me of "remedies" for this problem aside from "Don't use Cox PH then"? My boss really wants me to use Cox PH.
If all else fails, what could be a possible alternative for Cox PH? Non parametric survival analysis techniques? Thank you very much.
