How the automation pyramid enables commercial-scale ATMP manufacturing

1. What is the automation pyramid? 

The automation pyramid organizes the many levels of automation that are possible in any complex manufacturing environment into a simplified visual system. The information technology layer establishes the base levels of the pyramid, supporting the operational technology layer above. For ATMP manufacturers, the principles behind this pyramid are key to establishing scalability and flexibility, despite the challenges of small batch sizes.

Each level of the automation pyramid is interconnected with the next, creating a single, streamlined system that’s mission-critical for a viable commercial-scale ATMP operation.

To understand the importance of automation in a commercial-scale ATMP facility, imagine its opposite: a manually-driven facility in which operators are responsible not only for every unit operation but for every data notation, as well. That data is siloed in standalone spreadsheets, making it vulnerable to human error and leaving manufacturers exposed to serious regulatory and operational risks while facing the immense overhead costs of running a facility that resists efficient scalability. 

The automation pyramid proposes a different type of ATMP facility. It enables flexibility, scalability, and ongoing compliance by harmonizing data and building efficiencies into the manufacturing continuum. For ATMP manufacturers handling many small product batches, each generating a huge volume of complex data, achieving this level of data automation and harmonization is important both from a quality perspective and in terms of bottom-line business operations. 

In fact, when we asked cell therapy manufacturers about their cost-reduction strategies as part of last year’s Horizons: Life Sciences report, nearly half pointed to automation and inline analytics as key initiatives. Nearly a fifth plan to leverage process closure as well—an automation-driven capability that reduces the need for costly human intervention during manufacturing.

2. What are the key benefits for ATMP manufacturers?

By applying the principles of the automation pyramid to the ATMP facility, manufacturers can reduce their regulatory risks and continuously improve their operations using real-time, data-driven insights. A commercial-scale ATMP manufacturing facility would have a very difficult time operating as a viable company without automation and facility digitalization, so that’s the number one benefit of the automation pyramid: business resilience as you scale. 

Beyond that mission-critical benefit, building a strong automation strategy around your ATMP manufacturing operation will give you two important advantages:  

a. Regulatory compliance

By using a robust automated system rather than manual operations to collect data and control who has access to it, you can ensure that your data meets the principles of ALCOA+—that is, it’s attributable, legible, contemporaneous, original, and accurate, as well as complete, consistent, enduring, and available. Building these principles into your automation pyramid from day one will help you comply with the regulations that govern commercial-scale ATMP manufacturing, namely the FDA’s 21 CFR Part 11 and the EU’s GMP Annex 11.  

b. Continuous improvement

For companies making established biologics, the manufacturing process is well characterized and the cause of a potential deviation is often apparent. 

ATMP manufacturers, working in an all-new area of medicine, don’t yet have the same level of predictability over their manufacturing outcomes. Add to that the high stakes of working with source material from patients who are often extremely unwell, and subjecting that material to countless process steps, instruments, consumables, and so on, and you’ve got a manufacturing system that requires close monitoring and continuous, proactive control. That makes it all the more important to implement automated systems capable of collecting and analyzing huge volumes of manufacturing data. Using advanced technologies such as artificial intelligence and machine learning, your system will generate real-time insights to help you understand which operational levers you can pull to optimize your process while achieving the same therapeutic benefits for patients.

Even if you’re not ready for these advanced technologies right now, automating your data collection process means you’ll have a wealth of historic data to feed an AI system later on, when you reach that point. This will help you unlock more value, more quickly from your future investments in AI technologies, positioning you for ongoing improvement as your company evolves.  

3. What is the hierarchy of the automation pyramid?

Though nuances exist between industries, the automation pyramid is typically divided into five levels. Automated components in the pyramid’s lower levels help to measure and control outcomes on the facility floor (these levels make up the pyramid’s “Operational Technology” layer). In its upper levels, automation is the key to tracking the manufacturing process and driving enterprise-level decisions (these levels make up the pyramid’s “Information Technology” layer).

The operational technology layer

Level 0: Sensors and actuators → Enabling measurements

This level encompasses everything from sensors designed to capture pressure, pH, temperature, and other basic data to advanced Process Analytical Technologies (PATs) capable of measuring spectroscopy, viable cell density, and more.

Level 1: Automation → Enabling control

At this level, a Distributed Control System (DCS) or another type of control system, such as a Programmable Logic Controller (PLC), receives data from Level 0 and uses it to trigger specific operations, like opening a valve or turning on a pump.

Level 2: Supervisory control and data acquisition (SCADA) → Enabling context

At this level, operators can query the status of equipment, manage alarm systems, scan for trends in data from the first level, and more. The goal is to understand what’s happening on the facility floor.

The crossover layer: operational and/or information technology

Level 3: The Manufacturing Execution System (MES) → Enabling operations

The MES is a centralized hub that documents and tracks everything involved in manufacturing a specific product—in this case, an ATMP. The detailed and proprietary processes, best practices, recipes, quality protocols, and other steps that control your automated process are stored here.

The information technology layer

Level 4: Enterprise resource planning (ERP) → Enabling business operations

This level is home to the documentation and processes that are critical to overhead business operations. Though our focus in this article is on levels 0 through 3, this level is important, too—particularly for ATMP manufacturers, who need to manage high-stakes scheduling requirements.

When a patient needs a cell therapy, for example, several operations are set in motion, and each must fit perfectly into the next. Patient-specific material must be collected at the bedside and transported to the manufacturing site, and the site must have available capacity to handle that material. Meanwhile, a healthcare team may need to administer a pre-treatment before the patient can receive the therapy.

In order to multiply this level of complexity across hundreds of patient-specific batches–and do it efficiently and reliably– manufacturers rely on insights which are pulled from preceding levels of the automation pyramid and consolidated in the ERP.

4. Automation pyramid levels 0 to 2: What is process automation?

As its name suggests, process automation is a manufacturing strategy that leverages automated equipment and software to replace, improve, or accelerate manual operations.

Because of their small batch sizes, ATMP manufacturers face the challenge of repeating the same unit operations at a much higher frequency than other biologics manufacturers, with far more inputs and outputs in play and a greater demand on manufacturing equipment as a result. This can lead to bottlenecks across the manufacturing lifecycle and hamper manufacturers’ ability to scale or change their process as new opportunities emerge. 

Establishing and automating a standard process platform that’s easy to maintain, designed for open connectivity, and flexible enough to evolve as your process changes will solve many of these challenges. 

Benefits of process automation: 

• Lower regulatory risk

Recent developments in robotics and other automated technologies are helping ATMP manufacturers achieve process closure where it wasn’t possible before. That means eliminating or reducing both the bioburden risks and the risk of human error inherent in manual, open processes. The result is a stronger compliance strategy. 

• Real-time batch releases and reduced lab testing

Automation opens the door to accurate in-line data collection via PATs and smart instrumentation, giving you the ability to monitor certain critical quality attributes (CQAs) and critical process parameters (CPPs) in real-time. 

If you have a clear understanding of the CQAs and CPPs that define your product and you have the right automation strategy in place, you may be able to register certain deviations and make adjustments before engaging the quality lab—an efficiency that could save you significant time and resources, and that may help you save whole batches that might otherwise have had to be discarded at a great cost to patients in critical need.

• Next-level standardization

Every unit operation levies a cost in terms of the programming and validation effort involved. The more you’re able to standardize and automate those unit operations, the lower that effort will be, helping you to pivot between batches or products with more speed, flexibility, and cost control. 

• Improved scalability and modularity

Many ATMP manufacturers rely on modularity, and they need the ability to scale quickly by adding or moving equipment as their business matures, batch numbers increase, and product pipelines expand. That level of flexibility is much easier to accommodate in a manufacturing environment that’s highly automated.

• A future-proof process

An early commitment to process automation will pay off long into the future, in part because it positions you to rapidly adopt new technologies as they become available (such as a sensor capable of inline microbiological testing). In a manually-driven process made up of standalone, custom-designed unit operations, this type of opportunistic tech expansion is tricky and can elevate your costs.

5. Automation pyramid levels 3 to 4: How does automation impact manufacturing?

An automated manufacturing approach puts the right data in the right place, in real-time. The result is a database of electronic batch records (EBRs) that is up-to-date and searchable, and the ability to analyze and understand every interdependency linking the plant floor, the quality lab, and the final products that leave your facility.

We are at the top of the automation pyramid now, where manufacturing and enterprise-level planning take place. Our focus here is on the Manufacturing Execution System (level 3), and the role of automation in streamlining and optimizing the ATMP facility. 

If you’re just beginning your transition out of the research lab, you may think it’s too early to consider your MES. The fact is, though, the earlier you adopt your MES, the faster you’ll move. You don’t necessarily have to invest a lot of money—simply documenting your procedures in the MES from day one could have a significant impact when it comes to your tech transfer and validation experience. 

Benefits of adopting your MES early and automating your manufacturing process:

• Strong track and trace, chain of custody, and cold chain monitoring

The vital importance of ensuring that the right product goes back to the right patient is top of mind for every ATMP manufacturer—and a strong automation approach makes certain that it happens, every time. It’s also useful for ensuring that products removed from cold storage stay within specifications based on how they’re handled, and it improves traceability between the plant floor and the quality lab.

• Ongoing regulatory compliance

When your manufacturing operation is automated, you’re essentially outsourcing responsibility for data collection to a controlled entity that does not get fatigued at the end of a long shift, does not have illegible handwriting, and does not make keystroke errors. You gain the certainty that the manufacturing data you rely on to demonstrate your compliance with cGMP standards is always up to date in the eBR—not saved to a USB drive or logged in a 500-page printed batch record.

• Improved process control

The consistency and quality of an ATMP manufacturer’s final products depend in part on their in-process controls (IPCs). In order to eliminate variability, for example, some ATMPs may need to pass an IPC test before proceeding to the next manufacturing step. An automated approach can help you meet these IPC requirements while eliminating bottlenecks in the overall manufacturing process.

• Advanced data analytics

By automating the collection and harmonization of data from across different sources, including your process data, lab testing data, and manufacturing data, you give your data science team the ability to build meaningful data models,  recognize trends, and draw connections that would elude teams working with a manually managed database. 

6. What type of data flows through the automation pyramid?

A data strategy that harmonizes and contextualizes different types of data enables ATMP manufacturers to unlock the full potential of the automation pyramid.

A commercial-scale ATMP facility generates huge volumes of data. There’s process data, of course—that’s data generated during the manufacturing process itself, such as temperature, pH, concentration, etcetera. That’s just one data type, though, in a vast array that includes (but is not limited to): 

• Metadata and contextual data
• Quality control data
• Lot numbers and consumables data
• Chain of custody and cold chain data
• Environmental monitoring data 
• And much more

By building a strong data strategy into your automation pyramid, your facility will operate not like a series of disconnected silos, but like a single organism, each limb moving with coordination and control. 

To get there, start by incorporating FAIR data principles into your data strategy. That means ensuring that your data is findable, accessible, interoperable, and retrievable. Next, get clear about what you need to measure in the first place. Some data is important from a quality and process control perspective, while some is important for maintaining and controlling the automated systems themselves. Deciding what you will need to measure, and developing a data strategy around those needs, will help you get the most from your automation initiative.

That’s especially true for manufacturers who start their journey toward data-driven automation at the clinical stage. This early planning will set you up for a smooth tech transfer as you scale, giving you better control over all aspects of your facility, from proactively managing your inventory to optimizing your production schedule. Given that ATMP manufacturers operate in a paradigm that’s very different from traditional biologics—one in which hundreds of batches might pass through the facility in a single week—this level of control is absolutely necessary.