Optimal Censoring Rules Identifying Latent Hazard Ratios in the Analysis of Progression-Free Survival

• Activity Number: 192
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Suman Redhu* and Das Purkayastha
• Companies: Novartis and Novartis
• Keywords: Progression free survival ; Censoring ; Latent hazard ; Oncology

Abstract

Progression-free survival (PFS) is increasingly used as a primary end-point in oncology clinical trials. However, it's very common to censor patients due to anti-cancer therapy, or premature discontinuation of randomized therapy due to toxicity, or events after two or more missed visits in the analysis of PFS data. Different censoring rules in a clinical trial setup have effect on estimated PFS.

The primary thrust of this paper is to develop optimal censoring rules with inherently existing latent hazard in the analysis of progression free survival. A standardized hazard component confounded within each censoring rule or combinations used will be estimated using a maximum likelihood estimation method. Impact of censoring rules will be evaluated within a theoretical framework with an empirical illustration.