Concomitant or Prior Medication Tables

The following is a table specification for an overall summary of concomitant medications. As a rule for this summary, a patient should be counted only once per medication. Denominators are defined as the sum of all patients who received the specified study therapy. 

The following example relies on more PROC SQL, a little DATA step programming, a few SAS macro variables, and a final PROC REPORT for table presentation. Here are the concomitant medication summary annotated SAS program, notes for the program, and the output. 

O Instead of using a DATA NULL step as was done in Program 5.4, this program uses PROC SQL to define the three macro variables, nl, n2, and n3, that are used in the column headings for Active, Placebo and Overall therapy groups. The macro variables also serve as denominators to be used for percentage calculations. 

0 This DATA step rearranges the counts data set created by PROC FREQ. The data set is essentially merged with itself three times in order to get each treatment into its proper column. A group variable is created to help separate the ANY MEDICATION row from the other true medications. Percentages are calculated, and the columns (coll-col3) are formatted as XXX (XXX%). Finally, the lastrec variable is created to help make a continuation flag in the PROC REPORT output. 

Created by PAGE BREAK HERE... Summary C t conmeds.sas on 18JAN2 005. Table 5.5 of Concomitant Medication —(Continued)  

---