Application of Quality by Design in API Process Development

Design of Experiments (DOE) is the most effective method to achieve product and pro­cess efficiency and optimization. Design of Experiment (DOE) studies can help develop process development knowledge by revealing relationships, including multifactorial interactions, between the variable inputs (e.g., component characteristics or processing parameters) and the resulting outputs (e.g., in-process material, intermediates, or the final product).

QbD should by definition involve DoE to identify critical interactions and noise. But there are many steps in establishing a Design Space for QbD in API Process Development. This presentation will focus on the actual calculation and determination of a Design Space with many factors. The ICH Q8 document gives an illustrative example of how to handle this for two factors and broadly describes how additional factors should be treated. In an example based presentation we will demonstrate how the design space can be calculated for multiple factors, from around 2-20 factors. This presentation will also illustrate the fundamental difference between a design space approach to quality and a process set-point with a maximum process window. With a proper DoE we will be able to estimate a Design Space for one or several responses. Through advanced optimization tools in combination with Monte Carlo Simulations, this presentation will show that design space can be interpreted as spaces or regions of factor settings where all result specifications are fulfilled and with low risk of failure. The final Design space will be presented as a 2 or 4D contour probability plot.