ABSTRACT
The Efficacy–Toxicity (EffTox) trade-off-based design, introduced by Peter F. Thall and J. D. Cook has three basic components. The trade-off contours are an explicit structure that reflects subjective risk–benefit considerations weighing the hope of Efficacy against the fear of Toxicity. When constructing any phase I–II design, it is essential that the outcomes be defined collaboratively by the physician and statistician in a manner that is appropriate for the particular trial at hand. The definitions of Efficacy and Toxicity may vary widely depending on the particular clinical setting. The EffTox design treats Toxicity and Efficacy as bivariate binary outcomes. This makes the design straightforward to implement, but it has some drawbacks. Efficacy is a qualitatively different endpoint from Toxicity. Unlike Toxicity, a significant number of patients in the trial may not respond to the treatment and never achieve Efficacy.
