Satisficing, Optimizing and Globally Optimizing Trial Designs

December 15, 2020

When designing clinical trials, biostatisticians and clinical development teams are often faced with a conundrum. Given the parameters of their clinical study, they usually begin with five or six possible design options and begin to explore the most promising ones. The likelihood is that none of these trials will be optimal designs. Rather, they meet certain criteria that are “good enough” at which point, clinical development teams might begin to lead one way or another.

Choosing from a set of trials that are “good enough” uses a decision-making process that social psychologists call satisficing. Thereafter, the most favorable of these are improved by adjustments (e.g. an earlier interim look, or a slightly higher sample size). From the perspective of the space of all possible designs, clinical development teams often find a local optimum.

What about the global optimum, the absolute best design for your trial?

Cytel Chief Scientific Officer Yannis Jemiai explores the search for the global optimum in a recent Cytel podcast. Click to hear.

Watch Podcast

For more information download:

Read Whitepaper