(203b) A Methodology Based On Advanced Process Modelling for the Assessment of Risk Within a Quality By Design (QbD) Framework | AIChE

(203b) A Methodology Based On Advanced Process Modelling for the Assessment of Risk Within a Quality By Design (QbD) Framework


Sorensen, E. - Presenter, University College London
Bracewell, D. G., University College London
Close, E. J., University College London

There is an increasing desire within the pharmaceutical industry to develop and operate processes following a Quality by Design (QbD) strategy, where quality is built into the product and process based on a high level of product knowledge and process understanding.  Key to this approach is the identification and management of risk, which is defined as the combination of the probability of occurrence of harm and the severity of that harm. In practice, for industrial processes, a significant source of risk comes from process variability during drug manufacture, i.e. uncertain process parameters, for instance variable feed stream compositions. However, the majority of tools currently used to assess and manage these risks are qualitative, e.g. failure mode effects analysis (FMEA) and fault tree analysis (FTA).

In this paper, a methodology based on advanced process modelling is presented that can quantitatively assess risk associated with process variability or uncertainty. Risk due to uncertain parameters is rapidly identified by measuring the impact that uncertain process parameters will have on the ability of the process to meet its objectives. The risk can then be systematically reduced based on mechanistic understanding of the process design space. The available design space is identified whilst accounting for current process uncertainty. If the design space is not adequately robust, the level of control on critical process parameters required to reduce the uncertainty, and therefore bring process robustness to an acceptable level, can be found.

 To demonstrate the approach, we consider industrial chromatographic bio separations. A validated mechanistic model combined with stochastic simulation is used to identify critical and non-critical process parameters and inlet variables. The level of control which must be placed upon the critical parameters and inlet variables to ensure product quality and a robust design space is then identified. The presented methodology is a useful tool for risk assessment and reduction, which can increase the consistency of product quality and process performance, and thus bring us closer to the full implementation of QbD with full realisation of its benefits.