Covariance Estimation of Favorable Proportion for Missing Dichotomous Data in a Multi-Visit Randomized Clinical Trial

• Activity Number: 395
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Siying Li*+ and Gary Koch
• Companies: University of North Carolina at Chapel Hill and University of North Carolina at Chapel Hill
• Keywords: clinical trial ; sensitivity analysis ; missing data ; longitudinal data ; covariance estimation

Abstract

This paper presents a closed form method for sensitivity analysis of a randomized multi-visit multi-center clinical trial that possibly has missing not at random (MNAR) dichotomous data. Counts of missing data are redistributed to the favorable outcomes probabilistically to adjust for possibly informative missing; adjusted proportion estimates as well as their closed form covariance estimates are provided. Treatment comparisons over time are addressed with adjustment for a stratification factor and/or baseline covariates. The parameter estimates are computed via weighted least square asymptotic regression through randomization based methods. Application of such sensitivity analyses are illustrated with an example.