Data Structure

We use the colorectal cancer data for illustration of the analysis for the survival-binary surrogacy setting. The true endpoint is overall survival and

FIGURE 12.32

Data structure for the survival-binary setting.

the surrogate endpoint is a two-category tumor response: patients with complete or partial response are considered as responders and patients with stable or progressive disease are considered non-responders.

The data for each subject appears in a single line in which time-to-event (surv) and censoring status (survind) are given to the true endpoint and the response status (binresp) is the surrogate endpoint. The unit for the analysis is the Trial; see Figure 12.32.

The SAS Macro %SURVBIN

The joint model discussed above is implemented in the macro %SURVBIN that, for the colorectal cancer data, is called as follows:

%SURVBIN(data=colorectal,true=Surv,trueind=survind, surrog=responder,trt=treat,center=center, trial=center,vars=,patid=patientid)

The macro's arguments were discussed in Section 12.2. The argument surrog=responder defines the variable for the surrogate endpoint (2=respon- der, 1=no-responder) and it is assumed to be a binary variable.