Pharmaceutical companies are locked in a perpetual race against time.

Although patents can provide a company intellectual property protection for twenty years or more, twelve years or more of that time will be spent turning the ideas embedded in an individual patent to a marketing product, leaving only a few years to recover the often billions spent in development. Industry experts estimate that it costs $2.5 billion to bring a new treatment to market. Not only that, pharmaceutical companies lose $1 million a day in potential revenue for every day spent in the development or regulatory approval process. Only one in 5,000 molecules created gets commercialized.

The production process is time-intensive and lead times can last 365 days, 228 of which are dedicated to drug substance production, 75 to drug product formulation and 41 to packaging. Inventories, including the time spent on storing raw materials, intermediates or at in-transit distribution centers, can last 250 days. Can those times be reduced? Yes. A substantial portion of the time, materials are simply waiting around like passengers on a delayed flight.

Manufacturing innovations through digital technologies implemented over the next decade, I believe, will constitute one of the signature achievements of the industry. Not only will these changes help reduce the manufacturing cost and time now required to bring needed treatments to market, but it will also help pave the way for advances such as personalized medicine. Potentially, it could also help partly mend the rift between the industry, the public and regulators over drug pricing by both lowering costs and providing greater insight into how the industry operates.

In some ways, the success achieved by the electronics industry provides a blueprint for how the pharmaceutical industry could evolve. In the early 1980s, the largest and most prominent technology companies often operated as self-contained monoliths. Semiconductor manufacturers both designed chips and manufactured them. “Real men own fabs,” proclaimed AMD CEO Jerry Sanders. Likewise, computer makers took pride in developing their components and middleware. Rising demand and the rising cost and complexity of manufacturing infrastructure, however, lead to new business models and standards.

TSMC pioneered the fabrication facility for hire while ARM began promoting the then revolutionary idea of licensing its intellectual property to other producers. Now, few chip makers own facilities and nearly all of them license IP to finish their designs. Contract manufacturers graduated from being “screwdriver shops” to leading brands, like Lenovo, in their own right. Specialization helped fuel a virtuous cycle that has fundamentally changed virtually every industry on the planet. Meanwhile, the ‘rack and stack’ method of producing data centers and supercomputers with identical and easily replaceable servers and networking equipment led to the complete collapse of traditional supercomputer designs from companies like IBM and NEC.

Change and innovation in life sciences and pharmaceuticals will not and likely cannot evolve in the same way. Safety, efficacy and product quality must and will outweigh cost-saving considerations. Likewise, pharmaceutical companies operate within a regulatory framework that is in many ways unmatched in other industries. Nonetheless, opportunities exist to leverage innovations to streamline costs and production processes that in the long run could lead to lower costs, accelerated innovation and new business models.

Modular manufacturing
Pharmaceutical manufacturing is not for the faint hearted. Production facilities, gleaming with stainless steel machinery, traditionally can take three to seven years to build, cost $200 million to $500 million and are often suited only for a few products. Some companies spend billions over the space of a few years just on production sites. Like the semiconductor industry, the high costs are driven by stringent manufacturing requirements—high-throughput, high-volume, high-yield production in a clean room environment with a near-perfect level of consistency—rising R&D costs and the growing sophistication of the toolsets necessary for bringing a new product to market. Economic cycles and unforeseeable competitive challenges further compound the risk.

In developing a platform for an Alzheimer’s drug called Aducanumab, Biogen decided to take a different approach. Rather than build a large, unified production facility to meet demand, it came up with a way to ferment materials in seven smaller, identical and modular facilities that could function independently of one another without disrupting quality or consistency. With modular facilities, commercial production could be accelerated while capital expenditures could be spread out over a larger number of years. The modular approach also provides a way to scale more easily over time. The essential shift was made possible by using real-time data to fine-tune fermentation; put another way, real-time data and algorithms that could take advantage of it made small-scale production equal to large-scale. 

Using real-time asset and temporal context data and a mix of 58 different interfaces, Biogen created a BES Integrated Solution to move its testing process to the shop floor through a review by exception process. Using predictive models that continue to evolve with experience, the solution will predict batch quality during the production process. Now, rather than waiting for two-to-three days for a lab result, supervisors have access to real-time reports and notifications that highlight quality and process issues. With the combination of time series data coupled with manufacturing context, they can address issues before the batch moves to the next step and make adjustments in real time rather than simply scrapping materials.

With the BES Integrated Solution, Biogen expects to reduce testing time by 50 percent, batch exceptions by 70 percent and batch review time by 70 percent. Overall, the project will help Biogen decrease costs, increase yield and ultimately expand production efforts.*

Likewise, innovation helps reduce time to market and capital investment where manufacturers are experimenting with plastic “single use” equipment. The container approach, where manufacturers are producing materials in modular, 40-foot containers, can reduce the capital cost of production, minimize supply chain challenges and reduce energy and water consumption. With containerized manufacturing, new facilities can be erected in an area with existing customers, thereby reducing shipping time, potential tariffs and costs.

Continuous manufacturing
While the pharmaceutical industry has traditionally leveraged batch manufacturing processes, advancements in technology are pushing the industry towards continuous manufacturing. Now, rather than stopping production after each step to test samples, store sensitive product or ship materials to another facility for completion, continuous manufacturing allows nonstop production within the same facility. With an integrated production line backed by real-time operational data, pharmaceutical companies are eliminating hold times between batches and mitigating risks associated with storing sensitive ingredients. While continuous manufacturing has the potential to help drug manufacturers gain an additional $50 billion in annual revenue, it does require seamless process integration and utmost control. To achieve this, operational data and automation will play an even greater role than in batch processes. With new tools and systems, manufacturers can remove time constraints and decrease lead times as well as potentially reduce energy and increase throughput.

Layered analytics will play a critical role in moving toward a continuous idea, with some algorithms focusing on in-process optimizations or adjustments while others will be tuned to optimizing the process as a whole by mining years, if not decades, worth of historical data from the same and/or similar products. “Analytics” in the industrial world today are often viewed as a single market or product category. In the future, differentiated segments will emerge which, in turn, will potentially increase transparency, accelerate adoption and, yes, lower adoption and acquisition costs.

Personalized medicine
Personalized medicine effectively involves producing a batch of materials specifically tuned to an individualized genome. To perform this economically, a manufacturer must be able to produce only a very small volume a time, with batches often measuring in sub-liters. Think of it as modular manufacturing scaled down one step further in size and volume. Patients benefit from personalization while producers benefit from a lower production footprint, but the challenges of maintaining a golden batch become correspondingly larger. Personalized medicine will require distributed manufacturing to produce materials closer to patients and will change existing processes and economics; patients will ultimately benefit from improved treatment.

More precise testing
With innovation will come improvements in the drug testing process. Parexel, which specializes in contract research for biopharmaceutical and medical device industries, is building a platform for gathering patient test data with wearables. Currently, patients visit doctors at prescribed times, leading to fewer, more haphazard samples and high patient dropout rates. But, with wearables, companies have access to a constant stream of data, painting a complete picture through continuous monitoring. However, data collection will be an issue, as the sheer volume will overwhelm existing networks. Isabel DeZegner of Parexel estimates that a typical test patient can generate 2.5GB of data every day. To capitalize on the opportunity while managing the influx of information, companies must develop edge-cloud strategies for storing, caching and transferring data. It will take time to achieve, but it will give manufacturers increased visibility into the testing process, allowing them to reduce time to market and testing costs. 

Better facility utilization and compliance
Facilities are a critical part of the drug manufacturing process. Precise pressure, humidity, temperature, and other settings ensure that batches don’t spoil, but with those metrics come building constraints. Traditionally, scientists have worked around these constraints rather than buildings being configured to best suit their needs. With the help of digital technologies and operational data, Roche Pharmaceuticals is changing building plans. In Roche’s Basel lab facility, all walls are removable, enabling them to change room sizes to accommodate new processes, machines, and requirements. The lab is equipped with a hygienic barrier and a backup corridor, which means they can internally reconstruct the building configuration within a few days.

While room configuration is highly dynamic, regulations are strict. Each room is equipped with a panel on the door that shows current room parameters and, given permissions, users can change the room settings to meet needs. By using operational data collected from sensors, Roche can ensure that room settings, such as lighting and temperature, will produce the desired batch outcome and determine if any environmental or configuration changes have impacted production results. All prior room scenarios are saved, enabling users to save customized room settings and quickly reproduce prior settings.

For Roche, this configurable solution is saving them the costs associated with building additional rooms to accommodate future changes. Not only that, scientists have utmost control of their processes and settings so they can make real-time changes to ensure batch success.

Contract manufacturing
Due to safety concerns and regulatory requirements, the pharmaceutical industry has been reticent to adopt contract manufacturing models. For Lonza and Eli Lilly, real-time insights into remote manufacturers proved to be the missing piece for maintaining control.

Lonza, a biotechnology and chemical company, operates 32 plants worldwide under its Lonza Specialty Ingredients division. Plants manufacture a wide range of products, from vitamins to jet fuel. The manufacturing network is diverse, complex and scattered across multiple locations. Digitizing operations was the only way to solve operational challenges and ensure control. After identifying pain points within each site, they developed the Productivity Improvement with Operations Technology philosophy, which became the roadmap for the company’s smart manufacturing goals.

One of the biggest pain points was that process engineers were spending nearly half their time collecting and analyzing data for optimization projects. In response, Lonza implemented a unified data management system across all plants, replacing the existing individual management systems located at each site. Using a combination of the OSIsoft PI System, Microsoft, SEEQ and Hans-Meyer-Engineering expertise, they created a digital data infrastructure with advanced analytics capabilities that made process data visible on a global scale using the cloud. Now, employees have advanced visibility into information they need to perform their daily tasks, while the company has a scalable solution that allows employees to tackle pain points at any location. Overall, the project has helped Lonza reduce costs by several percentage points.

For 140 years, Eli Lilly has worked to help patients everywhere lead normal lives. A leading manufacturer of insulin pens, Eli Lilly works with a network of contract manufacturers to produce the pens. Made up of four-to-five parts, each pen is built using a series of molds and machines, all of which are owned by Eli Lilly. While contract manufacturing organizations (CMOs) have enabled the company to scale pen production, the remote nature of these facilities means lack of visibility and control. However, with the help of real-time sensor data and the cloud, Eli Lilly was able to stream real-time data from its contract manufacturing sites through the cloud to ensure equipment is running to specification and identify any production issues. With the influx of real-time data, Eli Lilly’s engineers can ensure that custom machinery is properly built and set benchmarks around machine performance before it is delivered to the CMO. Once delivered, they can see production data in near real-time, while it would previously take at least a week to view data. With a digital twin of the production floor, engineers can visualize machine downtime and drill down to understand the root cause.

Not only does access to real-time data give Eli Lilly confidence in remote operations, but these insights also allow the contract manufacturer to demonstrate accountability for quality and compliance as well as better control the supply chain—all while reducing production lead time and inventories. All in all, real-time operational insights means both pharmaceutical companies and contract manufacturers can meet regulatory requirements while ensuring high production standards.

Conclusion
Pharmaceutical manufacturers are at a critical juncture in history. Demand for innovative health care solutions continues to rise, but so do calls for new regulations and lower prices. The tension between investors, manufacturers, the public and the laws of chemistry cannot continue on its current trajectory. Innovations in manufacturing hold the potential to reduce some of the pressure by streamlining processes and costs that to date have been intractable facts of life. There may never be a Moore’s Law for pharmaceuticals, but some of the same concepts apply.

*Note: Biogen has halted work on Aducanumab, but continues to build the Solothurn, Switzerland facility due to come online in late 2020.

---

Petter Moree is Global Industry Principal, Life Sciences, Food & Beverages, Specialty Chemicals, at OSIsoft, LLC. Petter has a M.Sc. in technical chemistry with a specialization towards chemometrics and data science. After his M.Sc. from Umea University, Sweden, Petter joined Umetrics in 1999, where he worked for 17 years with customers in various markets for example Pharmaceutical, Pulp & Paper, Automotive, Semicon and chemical and with applications such as PAT, QbD, early fault detection, predictive analytics, MPC, and CPV. Petter was Product Manager for Umetrics’ portfolio, including SIMCA, SIMCA-online, and MODDE, and later became their global sales director.