Output Tab Descriptions | Clinical | Dictionary-Derived Term Results

Dictionary-Derived Term Results
Note: The name of this tab reflects whatever term is selected as Term Level. If Perform Double FDR Adjustment is checked, the terms for each statistically significant group and time window (if present) that appear on the Body System or Organ Class Results tab are presented. If Perform Double FDR Adjustment is not checked, all adverse events at the Term Level are presented.
The Dictionary-Derived Term Results tab contains the following elements:
 • One Bubble Plot of -log10(Raw p-value) by Risk Difference.
This Volcano Plot displays adverse events summarized at the selected Group Level by Trial Time Windows (if specified). The x-axis is chosen from x-Axis for Volcano Plot. In this example, the plot uses the difference in proportions between the treatments (Risk Difference). Other options include the log2(Relative Risk), which represents a doubling of the event rate for every one unit of change on the x-axis, or the log2(Odds Ratio), which represents a doubling of the odds of an event for every one unit of change on the x-axis.
The y-axis represents the -log10(Raw p-value). To interpret this axis, consider that
1. a p-value of 1 equals 0 on the -log10 scale,
2. a p-value of 0.1 equals 1 on the -log10 scale,
3. a p-value of 0.01 equals 2 on the -log10 scale,
4. a p-value of 0.001 equals 3 on the -log10 scale,
5. a p-value of 0.0001 equals 4 on the -log10 scale.
In short, the smaller the p-value, the larger the number on the y-axis (y can be thought of as the number of decimal places or number of zeros). Adverse events that are considered statistically significant while adjusting for multiple comparisons are above the dashed red line. This line is determined based on the selected Multiple Testing Method. The testing method considers the adverse event Group Level if Perform Double FDR Adjustment is checked. The p-value is from a Cochran-Mantel-Haenszel exact test. If Study ID varies among the subjects for analysis, the test is stratified by Study ID.
Bubble size is an indicator of the total number of subjects experiencing the event. Because numerous adverse events could be represented at the Group Level, the most statistically significant individual term defined at the Term Level within each Group Level by Trial Time Windows (if specified) is presented on this tab. Because Trial Time Windows are defined, a Group Level bubble is presented separately for each time window.
 • One Data Filter.
By default, Group Level and Trial Time Windows (if specified) are presented.