A Taxonomy of Estimands for Regulatory Clinical Trials with Discontinuations

• Activity Number: 319
• Type: Topic Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Thomas Permutt*+
• Companies: FDA
• Keywords: intent to treat ; missing data ; causal inference ; direct effects

Abstract

When patients in trials discontinue treatment because of toxicity, lack of efficacy, or death, it is not always clear how to define the effect of treatment. We might follow them, if they are alive, and record the outcome notwithstanding discontinuation of treatment. We might (and must, for death) consider the discontinuation itself to be the outcome. We might (rarely) consider the values before discontinuation to be the outcome. We might try to estimate the effect in a subset of patients who complete the course of treatment, though this requires careful definition. I shall talk about how to define and estimate these effects. I shall also consider and dismiss certain other definitions of effect, notably the effect if the patients who discontinued had not done so.