Treatment or Comparison Variable to Use

Use the drop-down menu to specify whether the data to be analyzed is based on the treatments actually received by the subjects or the treatment that was originally planned.

Available options are listed in the following table. Note these options may vary, depending on the specific report.

Selection

Description

Actual

Select this option when the actual treatments the subjects received deviated from the treatments that were planned and your analysis is of the data as treated.
Selecting Actual directs this report to use the values contained in the conditional TRTxxA variable from ADSL for the analysis.
If the TRTxxA variable is not found in ADSL, the report attempts to use the ACTARM variable found in the DM domain of SDTM. Failing that, the report attempts to use the TRTxxP variable and finally the ARM variable in DM.

Planned

Select Planned when the actual treatments the subjects received exactly match the treatments that were planned.
Selecting Planned directs this report to use the values contained in the required TRTxxP variable from ADSL for the analysis.
If the TRTxxP variable is not found in ADSL, the report uses the values in the ARM variable found in DM.

Specified Below

Select this option when the treatment is specified by a variable other than the conditional TRTxxA variable or the TRTxxP variable.

Best

Note: This option is available only in the Oncology reports.

Select this option to direct the report to take the best overall response recorded in the RS domain for each subject based on RECIST criteria terms for Complete Response (CR), Partial Response (PR), Stable Disease (SD), or Progressive Disease (PD).
If such values or tests are not found, the subject is marked as Not Evaluable or Not Recorded.

None

Select this option to plot the data without first segregating it by treatment variable.

Note: This option is not available in all reports.

For many Findings graphical processes, where the Treatment Variable is used for coloring the points, this option can be specified to avoid using any treatment variable.