2019 PharmSci 360
Models for dose-response based on longitudinal data combined with physiological and pharmacological principles can provide information for development decisions at a fraction of the study size required for empirical analyses based on end-of-treatment data. This comes at another cost, one of utilizing previous knowledge, assumptions and model misspecification. Minimizing this cost has been the objective of new modeling strategies. Model averaging is such a strategy which has shown attractive properties and found its way into mainstream drug development. Established model averaging and new approaches are the topic of this presentation.