Survival Analysis: Gastric Cancer, the Treatment of Heroin Addicts and Heart Transplants
But ‘‘survival times’’ do not always involve the endpoint death. This is so for the second data set that we will consider in this chapter; part of the data is shown in Table 15.2. Given in this display are the times that heroin addicts remained in a clinic for methadone maintenance treatment. Here the endpoint of interest is not death, but termination of treatment. Some subjects were still in the clinic at the time these data were recorded, and this is indicated by the variable status, which is equal to 1 if the person had departed from the clinic on completion of treatment and 0 otherwise. Possible explanatory variables for time to complete treatment are
maximum methadone dose, whether the addict had a criminal record and the clinic in which the addict was being treated. (These data are given in Table 354 of SDS.)
Part of the third data set to be considered in this chapter is shown in Table 15.3. Here survival time of potential heart transplant recipients is from their date of acceptance into the Stanford heart transplant programme. With these data, a patient is part of the control group until a suitable donor is located and transplantation is carried out, at which time the patient joins the treatment group. In this example, treatment is a time-varying covariate. Apart from treatment, the other covariates of interest are age (in years minus 48), whether the patient had previous
heart surgery and the waiting time for acceptance into the programme (years since October 1, 1967).