From automated to smart

Transforming life sciences manufacturing data into actionable knowledge

Julie Fraser
21 June 2017

6 min read

Despite being highly automated in many of their operations, most life sciences companies have not embraced smart manufacturing. Bridging this gap, experts say, will help life sciences companies control costs and be more responsive to patient needs.

With health care payers such as insurance companies and national health systems carefully screening what costs they reimburse, and new therapies serving ever-smaller patient populations, pharmaceutical, biotech and medical device manufacturers are looking toward the data generated by their processes for clues to capturing efficiencies.

“Most life sciences organizations are already collecting loads of data from automated manufacturing systems,” said Daniel R. Matlis, president of Axendia, a life sciences analyst and strategic advisory firm based in greater Philadelphia, with offices worldwide. “But collecting data is not enough; the value lies in extracting and distilling the data into information and knowledge to support smarter decisions.”

Benefits can include higher product quality, reductions in defects, as well as enhanced operational effectiveness and improved yields, Matlis said. “Gleaning intelligence from data can ultimately drive the shift from reactive to predictive approaches to product quality, which in turn improves patient outcomes.”

Smart manufacturing, Matlis said, requires automation in routine production and quality operations to generate data, plus ongoing analytics that support rapid and intelligent decision making. In life sciences businesses, much of the data for smart analytics comes from networks of manufacturing devices enabled for the Internet of Things (IoT).


”IoT is an enabler to improve product quality and increase production system throughput,” said Dennis Brandl, founder of BR&L Consulting, a Cary, North Carolina, firm that specializes in manufacturing IT, operations technology and industrial automation.

However, the life sciences industry has been cautious in adopting digital technologies, in part due to fears about changing a validated process that regulators have approved.

“When a bio-pharmaceutical product is submitted to regulators for approval, the chemistry and manufacturing controls are locked down,” Matlis said. “Changes that materially alter the manufacturing process must then be resubmitted for approval by regulators. As a result, companies often avoid making changes to manufacturing processes during the exclusivity period out of fear that the regulator might come back and push back on how they are doing it now.”

And yet, many aspects of digital measurement do not change the validated process itself.

“Putting sensors on a machine to monitor operations and time in operation is not really changing the machine or the validated production process per se, ” said Marcus Ehrhardt, a New York City-based principal at global consulting firm PwC’s Strategy&. “Given this fact, companies can take many steps toward operations digitalization without re-registering a product.”

The benefits, Ehrhardt said, are well worth the effort.

“By applying digital technology, companies can significantly increase visibility into their supply chain operations and make better and faster decisions,” Ehrhardt said. “Digitalization allows companies to fully integrate their supply chains and improve operational processes, making them more adaptive and responsive. As a result, planning accuracy, manufacturing efficiency and productivity, inventory levels, and service levels improve.”


To make the data meaningful, however, life sciences companies must understand it in the context of the processes that generated it.

“Process industries have not been strong at collecting the context of the data,” Brandl said. “Data can tell you the temperature, but not in relationship to what product or batch. Context for the data you’re collecting with the Industrial Internet of Things in the factory needs to be there while they are collecting the data – not recording it long after – or it doesn’t match FDA’s data integrity needs. Part of that gap is due to a disconnect caused by not closing the loop between operational information systems such as ERP and MES and automation systems that are actually collecting data.”

Complex supply chains in the life sciences contribute to the challenge, Matlis said, but well-managed data can help to solve it.

“Global and outsourced production processes further heighten the challenge in creating context to analyze production data,” Matlis said. “This is where IoT-enabled production, supported by information systems that have special logic to manage this decentralized activity, becomes essential. Smart and connected systems must support deep visibility control and collaboration across internal and outsourced processes, along with the data needed to support decisions. Smart operations also can be more flexible for manufacturing low-volume, specialized products at a reasonable cost, providing the ability to address unmet medical needs.”

Medical device manufacturers are increasingly analyzing the data generated by their processes to help them run a more efficient and profitable business. (Image © Bloomberg / Getty Images)


Supply chain complexity is causing some pharmaceutical companies to move toward continuous processing, where a single facility performs every step from ingredients to packaging. This trend may help speed the move to smart manufacturing because it will require replacing current production processes and manufacturing lines.

“We’ve developed not only a set of technologies that can help move the industry forward toward continuous manufacturing, but also a strategic approach, including collaborations between industry, regulatory bodies, and academics,” said Bernhardt L. Trout, director of the Novartis-MIT Center for Continuous Manufacturing (CCM) in Cambridge, Massachusetts. “Continuous manufacturing involves integration and an automated control systems approach. While there is still tremendous opportunity for new technology development, industry has been slow to integrate existing technologies. We think that this is changing for the better.”

Continuous processing allows for dramatic efficiency gains. For example, the CCM developed processes for the complete processing of active pharmaceutical ingredients (APIs) into coated tablets with fewer than half the steps of batch processing. This breakthrough eliminates the complex supply chains involved in producing the APIs, often in multiple plants, then sending them to a separate plant to combine with inactive ingredients that are processed into tablets or put in capsules, and yet another facility for packaging and shipment.

“Another benefit of continuous manufacturing is that you can select chemical technologies to increase throughput and yield,” Trout said. “In one example, we went from a couple of hours and 93% yield to five minutes and 98% yield. There are similar opportunities to streamline downstream drug product manufacturing.”

Rich data that is well analyzed is critical to identifying such efficiencies. Continuous processing makes it easier to benefit from the data, because processes can be optimized end-to-end.


Medical device companies, meanwhile, see an IoT ecosystem as a way to offer more value, creating information-based services that go well beyond selling devices.

BIOMODEX, based in Paris and Boston, for example, is linking scanned patient information to regulator-approved 3D modeling programs. The data is then used to drive its manufacturing process, which makes replicas of organs for surgeons to practice on prior to a difficult surgery.

“The process is quite simple,” BIOMODEX CEO and co-founder Thomas Marchand said. “We go from the medical image, such as a CT scan or MRI. Using FDA-approved software, then we go through a digital path to a 3D model ready for any multi-material 3D printer. Our blend of algorithms and materials science creates models and products that will react exactly the same as the patient’s body, even if both hard and soft tissues are involved.”

IoT-enabled connections between design and manufacturing allow this process to occur on-demand.

“Initiatives based on virtual human models, like The Living Heart Project, permit device makers to design and test their products in silico, accelerating time to market and driving improved patient outcomes,” Matlis said. “Another valuable use of virtual models is creating manufacturing simulations. This approach would enable companies, big and small, to use scientifically accurate digital 3D models to predict how production might occur – before sinking millions of dollars into factories and physical prototypes.”


The evolution to data-driven manufacturing processes in life sciences will not happen overnight, but the transition has begun.

“New technologies are starting to be adopted,” the Novartis-MIT Center’s Trout said. “We are not at a point where we have plug-and-play technologies for continuous or smart manufacturing. That is yet to come. Our vision is to integrate the entire system, end-to-end. The realistic way to implement it now is piecemeal, which is what companies have been doing.”

The biggest challenge, Axendia’s Matlis believes, is to create an environment for action. As a member of FDA’s “Case for Quality” initiative, he has a strong vantage point on the challenge.

“In addition to digital transformation, we need to change industry’s regulatory inertia,” Matlis said. “We are increasing collaboration between regulators, manufacturers, providers and other constituents to shift the industry’s focus from compliance to improved quality. This shift rests on quality metrics and will require embracing smarter manufacturing.”


While pharmaceutical, biotech and medical devices are very different, all can benefit from some of the same digital advantages:

Modular & mobile.Getting close to the patient is becoming a top priority to keep wait times short and opportunities for error low. This extends beyond medical devices custom 3D-printed at a physician’s office to creating modular biopharmaceutical production “pods.” Pfizer, for example, has a portable, continuous, miniature and modular prototype going into production. This is only possible with an integrated production information infrastructure that can ensure quality and compliance in a distributed environment through greater data collection for analysis and decision-making.

Connected. Increasingly, production equipment and products are becoming IoT-enabled and connected. As new processes are validated to include these smart production lines, factories can become more autonomous. Benefits from the IoT-enabled production equipment and products include allowing regulatory compliance records to be captured, processed and filed digitally.

Flexible. For better patient outcomes, companies must serve smaller populations cost-effectively. Whether producing small volumes inside traditional factories or using disposable production equipment, information systems must be capable of tracking and analyzing data on many products.

Consistent. To ensure consistency, a digital model should be used in production; ideally, that model will be based on the same data accumulated from early discovery through development, clinical trials and manufacturing. 3D-printed medical devices and patient-specific models benefit from consistent data as well.

Reliable. “Digitalization can lead to a transparency that helps to optimize manufacturing and maintenance while reducing downtime,” said Marcus Ehrhardt of PwC’s Strategy&. Machine-to-machine communication and machine-learning algorithms allow for seamless processes, predictive servicing of equipment and automatic corrective actions.

To discover how to improve the life sciences manufacturing process, visit:

Related resources