Creating Clinical Trial Summary Tables Containing P-Values: A Practical Approach Using Standard SAS Macros

P-value is a key criterion for evaluating the effectiveness and safety of new drugs in clinical trials, particularly in comparative studies. However, p-values are generally not presented in data summary tables generated with SAS software, because of the complexity of incorporating p-values into a formatted table that contains summary statistics, such as mean, proportion, or standard deviation. The purpose of this paper is to present a practical approach to create summary tables containing p-values which are calculated using customized SAS macros developed at STATPROBE, Inc. Three major types of SAS macros are presented in this paper to calculate p-value, according to whether variables are categorical (Fisher's exact test, Cochran-Mantel-Haenzel procedure, or logistic regression), quantitative (t-test, Wilcoxon rank-sum test, or general linear models), or survival (parametric or nonparametric estimates). SAS statistical procedures, such as PROC FREQ, PROC NPAR1WAY, PROC CATMOD, PROC GLM, and PROC LIFETEST, provide the function part of the macros. The output locations of p-values from SAS statistical procedures are also considered for the purpose of effective creation of the macros and generation of the summary tables. The programming techniques to prepare the input data set and to merge data sets containing p-values and other summary statistics are discussed with examples.