A0430
Title: Split-plot experiments with replicated runs in pharmaceutical synthesis
Authors: Martin Otava - Johnson & Johnson (Czech Republic) [presenting]
Kalliopi Mylona - King's College London (United Kingdom)
Abstract: Restricted randomized designs are essential in the pharmaceutical synthesis due to the operational restrictions and their cost-effectiveness compared to entirely randomized designs. Specifically, the split-plot designs are very effective in reducing the cost of an experiment in the presence of hard-to-change factors and/or of multi-stage processes. In classical designs, replicated runs for pure-error estimation are commonly employed, but they are rarely used in the more complex setting of restricted randomization. The reason is that, in practice, experiments in industry can rarely follow all the assumptions/conditions that are included in the methodological papers. A split-plot design based on a Bayesian D-optimality criterion can be adapted to ensure more precise pure-error estimation of the variance components and to fit the additional needs, the speedy implementation, and the restrictions that a real case scenario in industry often imposes. Emphasis is placed on the practical aspect of how to modify the complex design in a way that keeps the desired qualities and on how to assess the impact of more arbitrary changes.
