Cytiva专家谈生物工艺可观测性：实时监测为何重要及PAT落地难点

Bioprocessing generates more data than ever before. From core process variables, such as temperature, pH and dissolved oxygen, to more advanced bioprocess sensor data, including measurements of metabolites, nutrients and even real-time indicators of critical quality attributes (CQAs)—the ability to monitor processes has improved significantly in recent years.

However, despite this progress, the bioprocessing industry still struggles to use this data in a way that consistently improves performance. Increasing complexity and volume of data has outpaced many organizations' ability to integrate and act on insights from this data.

"The issue is not simply access to data, but how effectively it is used to support decision-making," says Ashley Howard, Senior Product Director for Automation and Digital at Cytiva, in a recent discussion.

Without real, usable insights into bioprocessing behavior, the ability to predict, control and improve performance remains sub-optimal.

Data exists—but it doesn't always translate into understanding

Most bioprocesses today are not short of information, but they are often short of insight. Data is collected across multiple systems and stages of the process, yet it is not always accessible, contextualized, or connected in a way that supports real-time data-driven insights and decision-making. In practice, this means teams continue to rely on traditional legacy systems—driven by delayed analysis, manual interpretation and subjective experience rather than real-time analytics for bioprocess control.

The Cytiva Biopharma Resilience Index, a survey that evaluates the biopharma industry's resilience across six key areas, reinforces this conclusion. It shows that while data generation continues to increase, organizations still struggle to translate data into timely, actionable decisions (1).

This manifests in familiar ways: variability is identified after it has already had an impact, deviations are investigated rather than prevented and decisions are made using incomplete or delayed information. As a result, batches are lost and processes operate below their full potential. In practice, while manufacturers can—and do—monitor processes, they are not always truly understood.

Steps toward improving PAT and real-time bioprocess monitoring

The ambition to improve observability is not new. The US Food and Drug Administration's (FDA) Process Analytical Technology (PAT) guidance, introduced in the early 2000s, was designed to encourage real-time process understanding and more proactive control strategies(2). In other words, PAT enables real-time monitoring of critical parameters and CQAs, which allows manufacturers to adjust processes proactively. This helps ensure product quality is achieved during the process rather than waiting until the end, when there may be a potential for batch loss if parameters are not met.

Since then, advances in bioprocess sensors and soft sensors, real-time analytics for bioprocess control and modeling have made deeper process understanding increasingly achievable. However, adoption has remained inconsistent, particularly in today's manufacturing plants.

As Ashley put it, "While widely discussed, PAT is not consistently embedded in practice. It is far harder to implement and costly at scale than many assume."

This highlights an important gap—not between what is possible and what is known, but between what works in theory and what is being realized in practice.

Why hasn't process observability in biomanufacturing scaled as easily as expected?

Process observability brings challenges in cost, validation, maintenance and regulatory scrutiny. Without a clear business case, organizations struggle to adopt and scale these technologies. This is not a technology gap, but an adoption gap.

Process observability is not just a technical decision

This is where process observability moves beyond being a technical consideration and becomes an operational one. In many cases, the limitation is not the availability of tools; it's the confidence to integrate and deploy them at scale within existing processes and regulatory constraints.

As with digital transformation more broadly, execution tends to break down at the point where organizations must commit to change. The value is understood. However, the path to implementation is less clear.

PAT and real-time bioprocess monitoring: from monitoring activity to enabling action

Process observability enables earlier detection of variability, faster decisions and a shift from reactive to proactive control. But it requires sensing, automation, data and process control to work together. As Ashley noted, “You can quickly see where systems are designed to work together and where they’ve been added over time.” Observability delivers more value when designed in, not layered on.

Observability must be designed in—not added later

Timing plays a critical role. Observability is not something that can be easily retrofitted. It needs to be considered early, during process development, equipment selection and facility design, so that it can scale effectively. If observability has been treated as something to be added later in the lifecycle, once processes are established, initial investments are made and manufacturing is underway—by that point, the cost of change is higher and flexibility is reduced.

What should have been a foundational capability becomes constrained by existing systems, validation requirements and operational realities.

Turning real-time bioprocess monitoring into a reality

If execution is the central challenge in modern bioprocessing, bioprocess observability is where that challenge becomes most visible. Without observability, processes remain reactive, variation is harder to manage and performance becomes less predictable. When observability is designed into process, manufacturers can be more proactive in controlling their process; this enables faster decisions and more confidence in their operations. However, making that shift is not about adding technology alone. It requires embedding bioprocess observability in a way that works within the realities of existing manufacturing operations. After all, in bioprocessing, the biggest challenge is not understanding what needs to be done—it is making it work at scale.

Read the full article here.

References

1. Cytiva. Global Biopharma Index 2025: The innovation illusion: why biopharma breakthroughs aren't moving the market. Published 2025. https://cdn.cytivalifesciences.com/api/public/content/HfzkmyqaSNSfYvvAVjLudQ-pdf. Accessed July 15, 2026.
2. US Food and Drug Administration. PAT—A framework for innovative pharmaceutical development, manufacturing and quality assurance: guidance for industry. Published October 2004. https://www.fda.gov/media/71012/download. Accessed July 15, 2026.