Abstract:
|
Multi-arm platform trials investigate multiple agents simultaneously, typically with staggered entry and exit of experimental treatment arms versus a shared control arm. Recent applications in the COVID-19 pandemic highlight the benefits of platform trials in evaluating treatments with operational speed and statistical efficiency. In such settings, there is considerable debate whether to limit analyses for a treatment arm to concurrently randomized control subjects, or to allow comparisons to both concurrent and non-concurrent (pooled) control subjects. The potential bias from temporal drift over time is at the core of this debate. We discuss the use of time-adjusted analyses (both Bayesian and frequentist) to model potential temporal drift, and the tradeoffs between bias and precision that should be considered in the evaluation of treatments with staggered entry.
|