The Pursuit of Digitisation: Process Centricity, Workforce Development and Ontology
19 October 2021
Automation & Digitisation
Skills, Talent & Development
In this Event Report, we explore the pursuit of digitisation, considering process-centric approaches, workforce and talent development, ontology and the five elements of the digitisation playbook discussed in the breakout rooms.
This meeting produced an exciting discussion among stakeholders in gene therapy and cell companies. Individuals present comprised a group with senior perspectives across all scales of operation, from start-ups in contract development and manufacturing to established CDMOs and academic experts.
Key subjects that fall under the scope of the SIG were discussed, including the pursuit of digitisation, process-centric approaches, workforce and talent development, and ontology. Lastly, breakout rooms allowed stakeholders to discuss and report back to the group on each of the five elements of the digitisation playbook.
“It’s critical for us to educate vendors. If they want their products used by more people, they need to provide the flexibility that product development needs.”
How do we pursue digitisation? Educating vendors on the desire for flexible product development (PD) is important; presently there is little flexibility in connecting to cloud-based data systems or due process development with different variables. A digitised system should support and benefit companies as they progress products through each stage of development and manufacture.
Clients desire flexibility in development and are keen to avoid being tied down with information systems used primarily for documentation purposes. Freedom and agility are critical, whatever the documentation system being used. In the development space, the barrier to entry is high. The sooner digitisation begins in a company, the better its subsequent flexibility will be.
Data architecture is common, but systems are not. Therefore, maintaining Critical Quality Attributes (CQAs) across systems and data sets can allow an organization to benefit from the improvements offered by data analysis as that data comes in.
The Process-Centric Approach
“Process orientation is extremely important; it’s how you move from development to commercial. Better yet, it empowers development to discover a greater understanding of their impact on commercial end products.”
How do we progress towards a process-centric approach to product development? Building a process model in PD can allow an organisation tobring to bear process-related automation to create batch records. Without this, an operation will invariably be experiment-centric or sample-centric, with data models never matching up to what is desired to push to manufacturing systems. And because off-the-shelf options for process-centric PD are limited currently, a risk exists for organisations wherein they may become dependent on a vendor or organisation. In-house development has merit.
Flexibility and focus are key priorities. Project work is mostly unique, meaning that data management often starts from scratch. Systems are made to standardise processes in facilities, but there is a lack of R&D process-focused systems. Manufacturing Execution Systems (MESs) tend to be inherently restrictive and encourage adherence to strict processes.
Workforce and Talent Development
“Culture and buy-in are fundamental. You can only go so far without it. It matters in every aspect of an organization from lab Q&A to technicians, IT, and more; the list goes on.”
Tech matters, but workforce is more important. To encourage digitisation at pace and scale, changes to academia and workforce development are vital. It must be a focus to empower staff to execute processes efficiently and to align themselves with the goals of an organisation.
Culture needs to shift back to academic translational institutions. Attempting digitisation across universities is a significant challenge and requires a drastic culture shift; faculty, grads, and post-docs never think this way. Until this shift is implemented at the academic stage of workforce development, efforts towards digitisation at scale will not translate into real careers and available future talent.
Using ontology structures to standardise data categorisation could prove powerful. Improvement is important, and discussion was had as to whether existing ontology could be standardised to support different facilities and development cycles.
With diverse stakeholders present in the SIG, breakout rooms allow for focused brainstorming on each of the five elements of the digitisation playbook.
End state vision: ‘No data shall be left behind.’ It’s vital that development and manufacturing are included from a data standpoint. We need to leverage standards in the marketplace and get vendors on board. Thinking long-term, we need enablement of tech like PAT and analytics and should consider a ‘Plug n Play’ concept of technologies versus broad integration efforts.
We must respect the fact that company maturity, financing, and more will vary; the path to digitisation will not be uniform across the industry.
Business case tools: What do we need to capture data within large-scale and early-stage development? We must deliver a portfolio of programs, projects, and technologies to advance CMC automation, with data gathered at all stages.
Tools and templates should empower organisations to ask the right questions and make their business cases. We also need to bring awareness to the reasoning and importance of data management as relates to regulatory requirements and business process and risk.
Digital and technologies: We need standardised data outputs. What happens to clinical data, and how do we reach a format where all elements talk to each other?
Gene and cell therapy are siloed; how do we get data out that we can use for digitisation? These fundamental aspects must be before explored we can gather enough data to produce digital twin-type activities before then going into pilot scale.
Roadmap: We need to consider long-term ICHQ12 guidelines, the realistic benefits of digitisation, and what that means post-market authorization. How, for instance, would an organisation change processes at a later stage? Our roadmap should support organisations of all sizes and should respect the need for incremental change in smaller operations.
The roadmap should include key elements such as ontology structures, communication between disparate systems, and capturing data within a given process. Lastly, we must consider data types generated in development and the data realistically captured during manufacturing; how can we link these two together for continuity, sampling, and data generation?
Workforce development: Cultural shift is our greatest challenge. How do we define new core skills and competencies, and how do we translate this to every stage of a new curriculum?
IT is now a fundamental enabler instead of an overhead, and there are significant gaps between traditional pharma and advanced cell and gene therapy. How do we change perspectives and fill these gaps? We should consider merging computer science and biotech and should facilitate cross-talk and collaboration between CDMOs and academics to help ensure fresh trainees are more prepared from day zero. Broad training time could be minimised, especially in the advanced therapy space, and managers can help to define ideal candidates.
The contract and research organization (CRO) AmplifyBio has announced the acquisition of select assets from privately held biopharmaceutical company PACT Pharma Inc to advance its cell and gene therapy service offerings.
For this Article, Phacilitate Editor, Georgi Makin, explores current challenges associated with CAR-T manufacturing workflows, considering how an integrated testing approach could reduce costs and accelerate speed to market.