Using SAS Proc Power to Perform Model-based Power Analysis for Clinical Pharmacology Studies

In this paper, we demonstrate that many hypothesis testing problems in clinical pharmacology studies can be reformulated as a oneor two-sample t-test or equivalence test. Thereby one can use SAS® Proc Power to perform the power analysis. Sample programs for different hypothesis testing problems are presented in a variety of study design settings. Compared to other commercial software such as NQuery and PASS, using Proc Power offers several advantages. First, one often obtains variance estimates by analyzing relevant clinical data in SAS, thus it is more streamlined and convenient to conduct power analysis in SAS as well. Second, the documentation in SAS is more comprehensive than those in NQuery and PASS; this helps the users better understand the procedure and hence the users are less likely to make mistakes. Third, as demonstrated in the examples, one can take advantage of the programming capabilities in SAS and make inputting parameters for power analysis more straightforward. Last, but not least, using SAS can better help the users keep a record of the power analysis, which is helpful for adapting the code for similar problems in the future.