Regulators and Industry Discuss Digital Transformation and Pharma 4.0™
The upcoming challenges and opportunities of digital transformation were the topic of a panel discussion held on 9 May 2023 at the 2023 ISPE Europe Annual Conference in Amsterdam, The Netherlands.
The panelists were:
The panel discussion was organized around the following themes:
Leading principles for technology and for regulatory oversight in digital transformation
The industry representatives emphasized the technical advantages of digitization, especially the visualization of data and the possibility of having process data available in real time. The Process Analytical Technology (PAT) concept can thus be fully realized. The automation of processes makes work easier. Nevertheless, digitization is not an end in itself and must not be viewed in isolation. Iacovelli stressed and Buendia concurred that digitization should be preceded by well-developed business cases that clearly define the purpose, target, and business value, then the “people aspects” of the change, and finally, the supporting technology.
The following were named as technical requirements: a common data layer, data standardization and harmonization, and the contextualization of data for data-based decision-making. Canzoneri commented that we should consider how to use pilot approaches. Overall, this can also lead to a harmonization of the way we view digitization as a business process.
The regulators emphasized the basic principles of GxP rules as well as the role of new technologies as enablers to improve GxP as well. The ICH Guidelines were referred to as a basis, for example, the need for management oversight as given in ICH Q10 (Pharmaceutical Quality System) as well as ICHQ3 (continuous manufacturing), which requires large amounts of data for monitoring. ICH Q9(R1) talks about digitalization and that new technologies should enable risk reduction. According to the basics of quality management "fit for intended use," new technologies should help to reduce risks, but should not themselves introduce new risks into the processes. As one panelist put it, "plug and play will not work by itself.”
New technologies should be understood for their true value, especially with regard to retained data, back up of data, traceability of data. "Just paperless is not sufficient” was a comment made.
Berends indicated that maintenance of systems requires careful consideration to ensure they remain in a state of control. Electronic systems should be more reliable.
As much larger amounts of data are generated and can be processed, it is a challenge to use, select and process all relevant data that can contribute to decision making.
Benefits and challenges for digital transformation
More data can be processed and analyzed faster. Correlations are better recognized and help with root cause analysis. Predictive analysis can help to avoid deviations. For new projects, proof of concept and proof of value can be calculated more precisely. Experimentation can be done virtually at lower cost and without risk.
In the greenfield approach, this is relatively simple; in the brownfield approach, specific business cases must be developed for old plants with old data architecture. Canzoneri warned that old technology can be a significant barrier and block implementation of increased digitization.
A particular challenge for the holistic approach of Pharma 4.0 is the creation of an end-to-end language across the functional silos in a company.
Another challenge is determining the required skill set and capabilities for the respective digitization project. IT knowledge and user knowledge must be managed. These organizational challenges can lead to a complete re-orientation of the organizational structure with a focus on interdisciplinary collaboration. Matrix structures or hybrid models are conceivable here. An exclusive focus on technology/IT will not lead to the goal.
Regulators again emphasized that companies often have not assured that data can be restored and could not demonstrate transparency. The industry representatives therefore also see a need to involve regulators in concept development at an early stage. Data integrity is, of course, imperative.
Furthermore, when correctly implemented, digitalization can reduce risk, and improve quality and compliance.
New technologies subject to regulation?
The dominant principle of risk management remains. Computers must not be allowed to make decisions; they must always be made by a human. It must be people who make decisions about batch release. Friedman stressed that human accountability for decisions associated with AI applications need to be clear. Machines should not be allowed to operate continuously without human checks at appropriate stages.
Proper standards for do's and don’ts should be known by people to ensure errors in designing AI/ML are avoided.
The practical question of how to deal with agile learning and the complexity of algorithms was generally answered by regulators that model evolution is subject to change management using a risk-based approach. Transparency and visibility of the vendors is vital with the need for sponsors to verify that vendors’ software is properly validated using challenge approaches.
Transparency also must be ensured in the learning process. Rosa referred to the FDA Medical Device Guideline of 2021 with regard to the existing regulations. Furthermore, guidance for the use of algorithms, e.g., sfor ML and AI, is expected in the update of the EC GMP Guide Annex 11, for which a Concept Paper was issued for comment earlier in 2023.
In summary, algorithms must be understood.
Cyber Security
IT and OT (operational technology) must be viewed holistically and combined in a security strategy. Corporate security and operational units must also be considered in combination. Especially important is the response time in case of security problems: red telephone hotline and a company-wide alert system must be implemented. Intensive training and education for all employees will become standard.
In conclusion, this new approach has implications for existing organizational structures and must also be seen as a subject of cultural change.
Disclaimer:
This is an informal summary of discussion panels held on 9 May at the 2023 ISPE European Annual Meeting. It has not been vetted by any of the agencies or regulators cited in this article, nor should it be considered the official positions of any of the agencies present.
