Smart, Real-Time Quality Insights Boost Life Sciences Manufacturing

View PDF

HTTPS://STOCK.ADOBE.COM

The COVID-19 pandemic has shone a light on restrictive business processes, information silos, and poor supply-chain visibility in many sectors. In biopharmaceutical manufacturing, for example, difficulties associated with product-quality management have been exposed and starkly felt.

However, public healthcare measures over the past 18 months have put physical distance between team members, thereby hampering the usual form-filling, manual sign-offs and spreadsheet-based recordkeeping associated with monitoring traditional manufacturing processes. In some cases, a lack of formal face-to-face discussions in the workplace that might have uncovered patterns of emerging problems have driven the need alternative means of problem solving. During the pandemic, the industry has been facing critical needs to implement effective monitoring, analysis, and reporting on product- and process-quality issues.

An Intelligent, Joined-Up View of Quality
Increased practical barriers to quality assurance and missed opportunities to identify and preempt issues along supply chains using data analytics have helped drive interest in intelligent, joined-up product- and process-quality monitoring. Such capabilities are based on a single, global, real-time graphical view of all aspects of production.

In the meantime, other parts of the biopharmaceutical industry have seen firsthand the benefits provided by preemptive signal detection (e.g., of an atypical result from a product-quality test). That is the case especially in pharmacovigilance. For such studies, the use of smart systems (e.g., digitalization solutions, real-time data, advanced data analytics software) provides a department’s best chance of accurately processing reams of incoming adverse-event data and meeting deadlines — with the confidence that nothing critical will be missed. Proactively monitoring and establishing alerts for potential manufacturing quality issues, product deviations, and/or process nonconformance would be other uses for such smart solutions.

The biopharmaceutical industry’s interest in harnessing smart, real-time product- and process-quality analytics is growing, especially regarding artificial intelligence (AI) and machine learning (ML). ML algorithms use historical data to find patterns so that they can spot the first sign of deviation and nonconformance in incoming data. Such capabilities can be applied to find solutions to different recurring problems such as those with equipment failures, variable impurity levels, product instability, and other issues whose causes need further investigation.

Quality Monitoring Goes Beyond Compliance
Traditionally, the tendency has been to view quality monitoring as a compliance activity, linked to the regulatory requirement for a periodic product quality review. That perspective, however, doesn’t take into account continuous, real-time product- and process-quality monitoring. The approach does not provide a chance to stave off production line issues before preventable risks and costs are incurred. If problems do surface while a biomanufacturer prepares for regulatory review, those issues are likely to be fairly established and would require investigation to determine their root causes, impacts to the process and product, and remedial actions required.

The traditional approach is a wasted opportunity, especially because much of the data to support continuous and timely quality tracking is being gathered anyway — with a view to creating that annual review report at some later date. But those data are not being amalgamated, compared, or processed in the moment to produce actionable insights and/or trigger alerts.

Drawing on Data from Across Functional or Departmental Silos
Moving to a situation that enables continuous, active product- and process-quality monitoring does not require a major upheaval but rather can be achieved in a series of small steps. The main objective is to establish systems that are capable of drawing on data from across functional or departmental silos so that deviation details, environmental data, compliance information, and corrective and preventive action (CAPA) records can be combined and cross-checked. Analytics and reporting tools should be able to call on historical data. That would enable biomanufacturers to compare data from multiple process runs. Such smart technologies also should be able to detect data deviates from current parameters immediately and compare those data to past patterns.

Biomanufacturers should maximize such capabilities by tailoring smart technologies closely to their specific circumstances. For example, a pharmacovigilance program for signal detection compares the number of occurrences of an individual adverse reaction with the total number of adverse reactions. A quality management program needs to be defined for each company, depending on the number of products, manufacturing sites, equipment, and other factors. The goal is to develop systems to address all those factors and create timely reports for each production process.

All biomanufacturers strive to be more effective and efficient with their resources, driving up product and process quality without over-extending internal resources. Smart, real-time monitoring technologies and reporting systems can realize such objectives.

A cloud-first, platform-based enterprise information technology (IT) environment will facilitate such capabilities. In some cases, when such initial systems are put into place, new capabilities and use cases can be added by switching on additional features or user groups. Those groups can draw on already-centralized, preintegrated data to tailor displays and applications for specific purposes.

Boosting Quality Delivers Immediate Returns
Most companies already have systems in place that capture and store data pertaining to processes, products, quality testing. Adding smart analytics and reporting capabilities can elicit immediate returns, thus saving resources, reducing waste, and ultimately preventing a substandard product batch from leaving the production line.

Rather than conducting quality reviews and investigations in hindsight, manufacturing teams can implement smart reporting systems to explore emerging issues and perform root-cause analyses in real time. If impurity levels exceed accepted norms, for example, teams can determine swiftly whether the probem might be a variance in the air humidity. This in turn might be traced back to a change of the heating, ventilation, and air conditioning (HVAC) system.

Providing access to real-time reports to the people or teams that need them is vital if prompt, effective action needs to be taken. Self-service analytics can be tailored so that people receive the reports that are relevant to their jobs and to ensure that teams can share and collaborate on data and analytics. As access control and collaboration requirements become more complex and granular, data security will be key to ensuring compliance.

Linking to Documents Underlying Data
To enable full automation of quality processes and analytics, process and product data and documents need to be integrated in one place rather than siloed. To provide quick reviews of underlying data, real-time notifications of issues must be linked to their sources. Having the ability to navigate to documents that underly reporting and visualization tools is important.

Manufacturing quality efforts focused on process and analytics also encompass high-quality document management systems. Document writing, review, approval, distribution, tracking, storage, and access all must be subjected to centralized quality processes to build a strong foundation for high quality in processes and analytics.

A central source of preconfigured documents is the starting point to support standardization to a preset level of quality. Quality documentation management should include management of standard operating procedures (SOPs) (encompassing policies, guidelines, and training documentation), quality control documentation, technical agreements, validation processes, and product quality reviews.

For the long term, getting quality relating to documentation correct not only supports better processes and analytics, but it also cuts costs by enabling content reuse for internal and external purposes. For example, product information can be reused in laboratories and in a technical agreement or contract.

Cloud-First Strategies Underpin Multiple Use Cases
The starting point for shifting emphasis toward continuous quality monitoring must be an amalgamation of data sources — ideally in a centralized, cloud-based repository that underpins multiple use cases. As long as contributing data (from analytics and process monitoring systems) continue to exist in spreadsheets and individual databases, the scope for viewing and acting on the complete, evolving picture (or any particular part thereof) will be limited severely.

Beyond using an integral master data source, configuring analytics, viewing, and reporting experiences to fit particular user requirements will be important. Equally important is the ability to set thresholds or parameters to trigger push notifications or alerts to the relevant people if such requirements are approached or exceeded.

The good news is that all of those capabilities are well within reach. Companies often can build on established data sources. Once responsible managers begin to think beyond compliance and toward efficiency and cost and risk reductions — benefits that come from being better informed — the business case for smart, real-time quality analytics in life sciences manufacturing tends to write itself.

Suggestions for Further Reading
Holowczak R, Furey J. Challenges and Benefits of Networking Process Control Manufacturing Systems: Integration to Business Systems in Industry 4.0. BioProcess Int. 18(5) 2020; https://bioprocessintl.com/sponsored-content/challenges-and-benefits-of-networking-process-control-manufacturing-systems-to-enterprise-resource-planning-systems.

Mleczko M, Maiti S, Spetsieris K. Establishing a Digital Platform for Data Science Applications in Biopharmaceutical Manufacturing. BioProcess Int. 20(1–2) 2022; https://bioprocessintl.com/2022/january-february-2022/establishing-a-digital-platform-for-data-science-applications-in-biopharmaceutical-manufacturing.

Morton J, O’Driscoll K. Designing the Right Strategy for Digital Transformation: How a Pragmatic Approach to Digital Transformation Can Help Biomanufacturers Adapt to a Challenging Future. BioProcess Int. 18(5) 2020; https://bioprocessintl.com/manufacturing/information-technology/designing-the-right-strategy-for-digital-transformation-how-a-pragmatic-approach-to-digital-transformation-can-help-biomanufacturers-adapt-to-a-challenging-future.

Rodriguez O, et al. Working with Big Data in Healthcare and BioProcessing Settings: A Brief Introduction to Key Components and Considerations. BioProcess Int. 19(11–12) supplement 2021; https://bioprocessintl.com/manufacturing/information-technology/working-with-big-data-in-healthcare-and-bioprocessing-settings.

Spetsieris K, et al. Pathogen Safety Digital Platform for Biopharmaceuticals: The Journey from Ground to Cloud. BioProcess Int. 19(11–12) supplement 2021; https://bioprocessintl.com/2021/november-december-2021-featured-report/pathogen-safety-digital-twin-case-study-for-biopharmaceuticals.

Siniša Belina is a senior life sciences consultant at Amplexor; sinisa.belina@amplexor.com; www.amplexor.com.