(470e) Model-Based Design of Experiments for Pharmaceutical Reaction Development

Optimal experimental design has long been a strategy for maximizing knowledge capture while minimizing experimental cost in many modeling applications for diverse fields such as geology, physics, biology, and economics. Where physical models are available to describe data, model-based design of experiments provides a framework for efficient experimentation to develop and refine model systems and their parameters. Pharmaceutical process development is no exception to the need for efficient experiments with limited material, resources, and time. In this work, model-based design of optimal experiments has been applied to systems of synthetic organic chemical reactions relevant to pharmaceutical process chemistry. Model discrimination techniques are used to discern potential mechanisms with targeted experimentation. Parameter identifiability and estimability analysis is applied to justify model structure and suggest optimal analytical methods. Parameter precision is maximized by predictions of sensitivity within traditional and novel objective functions applied to parallel experimental design. These model-based optimization strategies are combined in a workflow that provides resource-efficient reaction model building in short development timelines.