EU23CEDEC1/SES205 - Application, Validation and Verification of End-to-End Procecess Digital Twins

End-to-end models enable risk informed predictions of final product quality as a function of any process parameter (PP) of individual unit operations. This is where the resulting quality can be directly compared to drugs substance/ drug product specifications, relevant for patient safety and efficacy. This is beneficial over the use of process models from individual unit operations as they only predict intermediate quality and do not allow for assessing final product quality.  When it comes to verification and validation, a central approach to rate applicability and credibility of models within the V&V40 framework is to compare the model including its uncertainty to acceptance limits. The presentation will show how end-to-end process models can (1) correctly account for sources of variability in uncertainty quantification and (2) how to overcome the necessity to define acceptance limits for intermediate process steps.