Bayesian Hierarchical Monotone Regression I-Splines for Dose-Response Modeling and Drug-Drug Interaction Analysis: Application to In Vitro Studies

• Activity Number: 168
• Type: Contributed
• Date/Time: Monday, August 1, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Violeta Hennessey*+ and Veera Baladandayuthapani and Gary Rosner
• Companies: Amgen Inc. and The University of Texas MD Anderson Cancer Center and The Johns Hopkins University
• Keywords: Median-effect principle ; Combination Index ; Emax model ; Bayesian Effect Interaction Index ; functional data analysis

Abstract

We developed a flexible nonparametric method for meta-analysis of independently repeated dose-response experiments. Under the assumption of a non-increasing (or non-decreasing) monotone dose-response relationship, we make use of monotone regression I-splines (Ramsay, 1988) for estimating the mean dose-response curve. The complexities of monotonicity constraint can easily be accommodated under our Bayesian framework. We incorporate the splines into a Bayesian hierarchial model to address variability between-experiments, within-experiment (between-replicates), and variability in the controls. Markov chain Monte Carlo (MCMC), as implemented in WinBUGS, is used to fit the model to the data and carry out posterior inference on quantities of interest (e.g., inhibitory concentrations, Loewe Interaction Index for drug-drug interaction analysis). In addition, we explore a decision rule to assess drug-drug interaction. We compare our approach to analysis using the conventional parametric Median-Effect Principle/Combination Index method (Chou and Talalay, 1984).