Retrospective Meta-Analyses for Phase I Studies

• Activity Number: 372
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Sarah Zohar* and Anand Vidyashankar and Jie Xu
• Companies: INSERM UMR 1138 and George Mason University and George Mason University
• Keywords: phase I ; dose-finding ; meta-analysis ; maximum tolerated dose

Abstract

Estimating the maximum tolerated dose (MTD) is one of the critical aspects of Phase I analyses. However, in practice due due to small sample size if Phase I studies, the estimates of MTD are highly unreliable due to excessive variability between trials. In this presentation we describe three methods to combine information between trials to obtained a pooled estimate of MTD. The methods are based on (i) averaging (ii) random effect modeling and (iii) weighted averaging. We provide precise conditions when each of these methods are applicable in real practice. Additionally, we also provide a new methodology to construct confidence intervals for the pooled MTD. We provide a theoretical justification of our methodology and illustrate our results using several examples.