Accounting for Pre-Treatment Dropout Using Inverse Probability Weighting: Application to a Phase III Clinical Trial in Multiple Sclerosis

• Activity Number: 350
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Stephen Lake*+ and Amy Cinar and Jeff Palmer and David Margolin and Michael Panzara
• Companies: Genzyme and Genzyme and Genzyme and Genzyme and Genzyme
• Keywords: dropout ; inverse probability weighting ; multiple sclerosis ; missing data ; proportional hazards

Abstract

In open label studies, patients may opt to discontinue prior to treatment after learning to which treatment they were randomized to receive. This patient selection process has implications for statistical inference if the patients who withdraw prior to treatment are different than the patients who are treated and/or if the withdrawal is imbalanced between treatment groups. The use of inverse probability weighting to account for dropout in this setting was explored. Simulations results demonstrate that this is an effective means for addressing this type of dropout. This approach was applied to an open label, rater-blinded, phase 3 clinical trial in multiple sclerosis. The results of a sensitivity analysis using inverse probability weighting demonstrate that the primary efficacy results from the study are robust to pre-treatment dropout.