Issue link: https://www.e-digitaleditions.com/i/1517675
www.biopharminternational.com Quality and Regulatory Sourcebook eBook March 2024 BioPharm International ® 15 Quality by Design cross-functional learning and highly informed decision making. Digital transformation of risk assessments can re- duce the time and human errors possible from traditional risk management, technology transfer, and reporting. A generalized QbD data flow for an example biophar- maceutical product using standard non-digitalized, un- integrated tools is presented in Figure 2. Prior knowledge and process development data may be stored in databases linked to laboratory systems and/or published as indi- vidual analyses and reports. The information from these sources' feeds directly into the product and process defi- nition, which can include process tech transfer (process description) documents, bill of materials, and the quality target product profile (QTPP, which may also include a list of critical quality attributes [CQAs]). The product and pro- cess definition (i.e., product attributes and process recipe) acts as the defining input for several QbD risk assessments, including a CQA risk assessment (which in turn produces a CQA document), various process risk assessments, and raw material risk assessments. Additional contextualiz- ing platform knowledge and process development data sources are also leveraged to guide the execution of these risk assessments. The main outputs of these risk assess- ments—often summarized at specific timepoints in indi- vidually published reports—are material, parameter, and attribute criticalities, along with critical process controls and information regarding process technical risks and mitigation plans. A parallel analytical risk assessment process links the analytical target profile (ATP) (9) to an- alytical procedure development with the identification of critical analytical procedure parameters (11). The out- puts of both process QbD risk assessments and analytical method/testing strategy information, along with product quality data obtained from clinical and/or commercial manufacturing, are used to document the overarching product controls strategy. The control strategy is then used as a key source of information for Module 3 of the common technical document (12) of the Biological Licensing Ap- plication (BLA) (or New Drug Application (NDA) for small molecules). Additional overlap and data flow between the regulatory filing and risk assessments is seen in the case where established conditions (ECs) are specified for a prod- uct per ICH Q12 (5). There are obvious drawbacks to this standard data flow. First, even this generalized and simplified view is seen to be highly complex. The view in Figure 2 also only reflects a single lifecycle iteration, and as such, this process would repeat multiple times through process development, tech transfer, and manufacture at clini- cal and commercial scales. Such recurrent complexity creates inherent inefficiencies, and hence increased time and cost, in completing key chemistry, manufac- turing, and controls (CMC) deliverables. Consequently, CMC project team members are forced to navigate the web of interconnected inputs and outputs to manually (and often painstakingly) search for the data needed to FIGURE 2. Legacy "paper-based" quality-by-design (QbD) data flow (generalized for an example biopharm product and simplified to a single lifecycle iteration).