A Case Study of a Complex Design for a Clinical Trial with Features of Randomized Withdrawal and Parallel Randomization in a Rare Disease Area

• Activity Number: 343 - Innovative Trial Designs and Analytics
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Junxiang Luo* and Qi Zhang and Hui Quan
• Companies: Sanofi-Aventis and Sanofi and Sanofi US
• Keywords: rare disease; inverse probability weighting; historical data borrowing

Abstract

Complex clinical trial designs for rare diseases raise statistical challenges. Treatment (trt) A is investigated to add on to the standard of care (SOC) with a two parts trial design. Part 1 includes two periods: first period is an open label trt A added on to SOC for 1 year to assess patients’ efficacy and safety, then the tolerated patients are re-randomized to SOC + trt A or SOC for another 1 year in Period 2 following a sufficient washout. Part 2 randomizes new patients to SOC + trt A or SOC for 1 year after completion of Period 1 of Part 1. In addition, the trial considers borrowing historical information from patients treated by SOC through an informative prior. A statistical approach is proposed to this complex design: 1) within patient correlation is considered for two periods' data in Part 1; 2) the inverse probability weight technique is used to adjust for the selection bias in analysis of combining data from Part 1 and Part 2, as only tolerated patients are eligible for Period 2 of Part 1; 3) a mixed or commensurate prior is implemented to borrow historical information. Simulations are conducted to compare the performances of the approach under different scenarios.