24/25 March 2020
Process controls have become customary in pharmaceutical processes. The aim is to monitor a process and improve it if possible. With changes to the FDA's Process Validation Guideline and Annex 15 of the EU GMP Guidelines, this process monitoring has also gained more regulatory importance as a part of the product validation lifecycle (Ongoing/Continued Process Verification). Statistical process control (SPC) is viewed as one option to implement Ongoing/Continued Process Verification. PharmEuropa, the publication medium of the European Directorate for the Quality of Medicines and Healthcare (EDQM), has now published a new draft chapter (5.28) about SPC for the European Pharmacopoeia.
This chapter particularly covers multivariate statistical process control (MSPC). To start with, however, the "classic" univariate SPC is being explained. In univariate SPC, a single parameter is being monitored with a control chart. The article describes how this control card is to be set up. The advantage of this kind of control cards is the relatively simple mathematics behind them.
In contrast, multivariate statistical process control can be used to monitor several, even correlating, parameters. This enables its use in the application of process analytical technologies (PAT), continuous manufacturing (CM) or real time release testing (RTRT). In principle, various univariate SPC control cards could also be used for different parameters; however, those parameters should not be correlating. If they do, an MSPC would be the better choice. The development, application and theoretical background of MSPC are outlined in the draft chapter. However, the mathematics behind MSPC is much more complicated than that behind univariate SPC.