Development and Implementation of a Quality Control Strategy for an Atropisomer Impurity Grounded in a Risk-Based Probabilistic Design Space

In the development process of an active pharmaceutical ingredient (API), determination of the design space used for control of critical quality attributes (CQAs) is a key component of the quality by design (QbD) framework outlined by the regulatory agencies. Herein, we propose the use of probabilistic (Bayesian) methods to drive the development of the design space and control strategy for a process. Using probabilistic methods to quantify the risk of failure for different processing options allows for informed process design and control strategy decisions, enabling robust processes. We present a case study of a complex API reaction and crystallization: first, probabilistic models are built using lab and plant data. Next, these models are used to compare different processing options in the context of reliability pertaining to a critical quality attribute (CQA). A process decision is outlined based on the reliability estimates from the models, and last, a control strategy is proposed for the CQA for a defined reliability specification. This case study highlights the use of probabilistic modeling as a tool for efficient and robust process design in the pharmaceutical industry.