Constructing Model for Survival Data by the Makeham-like Unproportional Hazard
Penulis/Author
Dr. Adhitya Ronnie Effendie, S.Si., M.Si., M.Sc. (1); Drs. Danardono, MPH., Ph.D. (2); Prof. Drs. Subanar, Ph.D. (3)
Tanggal/Date
2009
Kata Kunci/Keyword
Abstrak/Abstract
In this research, we propose a new method in model construction of survival data. This research was originally based on Cox regression, or more popular with Cox proportional hazard. The main idea of the research is we thought that in Cox (1972), the model with exponential function, h t h t ( | ) ( )exp( ) Z Z = 0 b¢ was a Gompertz-like proportional hazard with drift parameter h0 , the baseline hazard rate. As we can predict, we can generalize the model using Makeham framework to get ( | ) ( ) ( ) exp( ) mak h t h t h t Z Z = + i d b ¢ with h t( ) is the intercept baseline hazard and h t( ) iis the drift baseline hazard. The result will become unproportional hazard because ( | )/ ( | *) mak mak h t h t Z Z will not a constant. This research will investigate the partial likelihood for distinct-event time data as a concequences of the new model
