A New Modeling Approach for Quantifying Expert Opinion in the Drug Discovery Process

• Activity Number: 30
• Type: Contributed
• Date/Time: Sunday, August 4, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Ariel Alonso and Elasma Milanzi*+ and Geert Molenberghs and Christophe Buyck and Luc Bijnens
• Companies: Maastricht University and I-Biostat, Hasselt University, Belgium and Universiteit Hasselt & Katholieke Universiteit Leuven and Janssen Pharmaceutica and Janssen Pharmaceutica
• Keywords: Selection bias ; Combined model ; Sensitivity

Abstract

Expert opinion plays a prominent role when selecting promising chemical compounds in the drug discovery process. However, when analyzing this information, complex hierarchical models often need to be fitted. Indeed, frequently one needs to explicitly state a model for the so-called selection process and then fit it jointly with the model for the ratings. We call the previous approach the joint modeling approach. Unfortunately, misspecifying the selection model in this paradigm may impact the estimation and inferences on parameters in the rating model, which are of most scientific interest. In the present work we incorporate the selection process into the analysis by adding a new set of random effects to the rating model and, in this way, avoid the need to model it parametrically. We refer to this approach as the "combined" model approach. Through simulations, the performance of the combined and joint models were compared in terms of bias and confidence interval coverage. The estimates from the combined model were nearly unbiased and the derived confidence intervals had coverage probability around 95\% in all the scenarios considered. In contrast, the estimates from the joint model