Predicting the Timing of Final Events in Blinded Clinical Trials

• Activity Number: 610
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Marc Sobel* and Ibrahim Turkoz
• Companies: Temple University and Janssen R&D
• Keywords: survival analysis ; beta and gamma processes ; Polya Tree models ; predictive inference ; Schizoaffective disorder

Abstract

Variable length time to event clinical trials are based on observing predetermined total numbers of events. Many of these trials involve following patients over a long span of time. The primary events of interest are e.g., death, relapse, adverse drug reaction, or new disease development. Research teams need to determine the conditions under which studies can be finalized in order to put together resources necessary to the dissemination of study results. We propose strategies for determining the timing of the final events in blinded settings. We employ nonparametric Bayesian survival models for this purpose. Each model operates under real world assumptions. One such model assumes that the hazard rates of subjects evolve as a function of the length of time since the beginning of the study. Another model assumes that the hazard rates of subjects evolve as a function of the length of time patients spend in the study. We employ nonparametric Bayesian frailty models to make predictions when the hazard rates between treated and non-treated subjects can be assumed to differ. Methodologies are tested with 334 subjects using a randomized, double-blind, placebo-controlled trial.