Modern Approaches to Clinical Trials Using SAS: Classical, Adaptive, and Bayesian Methods

4.7 SAS MACRO BSMED

4.7.1 Implementation of BSMED

4.7.1.1 Detailed Macro Options

In this section, we present the details of SAS macro, BSMED, for Bayesian survival meta-experimental design. As discussed in 4.2 Meta Design Setting, the macro BSMED requires two input data sets, namely, the current data and the historical data (optional), 11 input variables, and one output file (optional). Names and Descriptions of Input Variables and Design Values shows the names and descriptions of the input variables and the design values in the SAS macro.

Table 4.4 Names and Descriptions of Input Variables and Design Values

Data	Names in macro	Description
Historical	a0	Discounting parameter
	sigma0	Initial variance of
	sigma1 ...