The schizophrenia study is used for illustration of the binary-binary setting. The true endpoint is based on the PANSS score and the surrogate endpoint is based on the CGI score. The binary endpoints PANSS/CGI reflect the presence or absence of clinically relevant change in schizophrenic symptomatology. Clinically relevant change is defined as a reduction of 20% or more in the PANSS scores, i.e., 20% reduction in post-treatment scores relative to baseline scores, or a change of 3 points in the original CGI scale (Kane et al., 1988; Leucht et al., 2005). Hence, the true and surrogate binary endpoints are defined, respectively, as:
A partial print of the data is given in Figure 12.45.
Data structure for the binary-binary setting.
Individual-level and trial-level surrogacy measures with 95% C.I.
The SAS Macro %BINBIN
The reduced fixed-effects model formulated in (12.47) can be fitted using the SAS macro %BINBIN, which has the general form
%BINBIN (data=schizo,true=cgi,surrog=panss,trt=trtmnt, trial=investid,patid=patientid,looa=1)
The macro's arguments were defined in Section 12.2.