Abstract:
|
The standard approach to investigating the effects of intervention/prevention regimens on continuous outcomes in randomized clinical trials (RCTs) is via the average (i.e., mean) causal effect. However, there is reason to believe that averages can hide important variation in the treatment effect in many real-world applications and can potentially be misleading when determining the appropriate treatment for a given individual. Thus, statistical methods which directly address such treatment effect variability (TEV) are needed. Although not widely used, there are statistical methods currently available to address TEV. However, it can be important to augment these statistical quantities with graphical representation to more fully explicate the nature of TEV within a given RCT. In the present talk, I discuss and illustrate the challenges underlying representing TEV graphically, and further demonstrate the clear benefit of such graphical approaches for providing a clearer understanding of the treatment effect in RCTs, when compared to approaches based on the average causal effect only. An illustrative example is used to demonstrate these benefits.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.