Global Sensitivity Analysis for Randomized Trials with Informative Assessment Times: A Fully Parametric Approach

Zebin Wang, Daniel Scharfstein

Abstract:

Many randomized trials are designed to collect outcomes at  xed points in times after randomized. In
practice, study participants miss their assessments or are assessed o -schedule. If and when individuals are
assessed may be informative. In this thesis, we develop a sensitivity analysis methodology for analyzing
randomized trials with a potentially informative assessment process. We develop these methods in the
context of the Asthma Research for the Community (ARC) trial.