Event Rate Estimation Based on Blinded Time-To-Event Data to Verify Sample Size Calculation Assumptions

• Activity Number: 243 - Contributed Poster Presentations: Biopharmaceutical Section
• Type: Contributed
• Date/Time: Monday, July 31, 2017 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Youlan Rao* and Suhyun Kang and Jody Cleveland and Chunqin Deng
• Companies: United Therapeutics Corporation, Durham, NC and North Carolina State University, Raleigh, NC and United Therapeutics Corporation, Durham, NC and United Therapeutics Corporation, Durham, NC
• Keywords: Event Rate Estimation ; Blinded Data ; Sample Size Calculation ; Hazard Ratio ; Pulmonary Arterial Hypertension

Abstract

For a randomized double-blinded two-arms clinical trial with time to event as primary endpoint, number of events required are derived based on log-rank test with assumed event rates in placebo and treatment groups. This paper presents several methods to validate the assumption of event rate for the placebo group based on blinded time to event data in interim analysis. We propose to perform a pooled blinded analysis after a portion of the required events of interest have been reported. A mixture of exponential distributions will be fit to the time to event data pooled across groups under the assumption of a fixed hazard ratio (HR). The placebo event rate, using maximum likelihood estimation, Bayesian approach and expectation-maximization (EM) algorithms, will be back calculated from the pooled data. Simulated data will be used to compare these methods in terms of bias and variance in event rate estimation. A hypothetical example in pulmonary arterial hypertension (PAH) trial with time to first morbidity/mortality event endpoint will be used to demonstrate the application of the methods.