From Industrial To Investment Strategies In The Life Science Sector
By Wayne Koberstein, Contributing Editor
Observations at the Emerson Global Exchange and BIO Investor Forum illustrate the great gulf between the plant floor and C-suites of biotech and Big Pharma.
If you could start today, building a big company from the ground up with nothing to hold you back, you might succeed in constructing a fully integrated, uniform, company-wide manufacturing operation equipped with all of the most advanced technology available. Your production facilities would all look and act the same, organized as a common quality-by-design (QbD) system, optimized by process analytical technology (PAT), updated regularly, and constantly calibrated with automation to achieve reliable and predictable output for all of your products.
But you can’t. And consequently not a single company on earth fits the preceding description. Big Pharma companies, nearly every one an amalgam of acquisitions and legacy systems, continue to resist sweeping changes in manufacturing, and life science investors share their avoidance of the manufacturing challenge and opportunity. That was the big lesson that returned to me over and over again at two quite different industry events I recently attended: The Emerson Global Exchange and the BIO Investor Forum, both held the same week in early October.
Emerson Global Exchange does not fit into the usual categories of industry conferences or exhibitions; Emerson is a huge industrial supply company serving many sectors, notably oil and gas and mining, with life sciences a relatively small but quickly growing slice of its business. The Exchange hosts the company’s clients, and almost all presentations and exhibits address its products and services, with some room given to industry overviews and partner suppliers. But I found the life sciences sessions quite useful, especially for seeing how biotech and pharma manufacturing appear at the plant and unit management levels.
BIO Investor Forum focuses primarily on venture capital, stocks, partnering, and licensing for biotech and pharma. With the large companies, investors, and analysts in the background, small companies present their dog-and-pony shows in 15 to 30 minute rounds along four simultaneous tracks over three days. I attended the last day, seeing seven presentations, and meeting with several companies privately. A plenary session at the end featured a panel of VCs giving their forecast for 2013. It was the proverbial 30,000-foot view of people who configure the industry — and its lowly functions, such as manufacturing — only in financial terms.
NEW INDUSTRIAL VISION FOR BIO & PHARMA
Few companies, like Emerson, encompass the manufacturing side of so many different industries, and, thus, the Exchange put biotech and pharmaceutical production in a much larger context than we are accustomed to seeing. To walk the floor of the exhibition was to cross the bridge of time from artifacts of the 19th century — huge cast valves standing tall over the crowd — to working models of the latest automated systems for real-time monitoring and control of complex processes. It was easy to recognize pharma/biopharma manufacturing in the former but not in the latter.
At quiet moments, at a table or lounge isolated from the throng of attendees, heads around me nodded when I voiced the thesis that C-level executives in bio and pharma rarely poke their noses into manufacturing. One person related: “When I became head of a large pharma company’s production unit, the CEO called me in, introduced himself, and then told me to make sure he didn’t see my face again. He said that to see me would mean there was a problem — and he didn’t like problems.”
A life sciences forum echoed that theme, if more delicately. Dubbed “Process Robustness. From Molecules to Medicine,” the forum began with some wise words from John Berra, retired chairman of Emerson Process Management, who stressed the potential of automation to boost manufacturing quality in the life sciences, where productivity, efficiency, and safety of pharma and biotech plants can have life-and-death consequences — and affect their exposure to regulatory and legal complications.
Automation in pharmaceuticals and biotech has succeeded in a vertical sense, he said; engineers have used traditional sensing, feedback, and control mechanisms effectively at various points along the production line. But such solutions tend to be unique to each factory and without linkages between the different production stages, from process development to clinical and commercial manufacturing. Horizontal discontinuities slow product development because they interrupt a potentially valuable flow of information and knowledge from the bench up to full-scale production and back again.
Scott Broadley of Broadley-James followed with a demonstration of how horizontal automation propels such information flow by creating “better tech-transfer packages” for process development, scale-up, and production. He described a project that teamed banks of smallscale bioreactors, networked and controlled by advanced industrial automation systems, and used to generate data for multivariate analysis predictive of large-scale process and output.
Immediate advantages of the automated banks over traditional scaleup modeling systems included more runs and greater throughput, more successful runs, operator efficiency, and the ability of the set-up to integrate with process-control technology. Such automated banks mimic process control strategies used at pilot and production and minimize data variance between bioreactors, batches, and facilities. Ideally, the scaled-up production system would follow not only the key process parameters produced by the small-scale model, but also — horizontally — its overall automation and bioprocess control approach.
FROM PROCESS TO PRODUCTION, TECHNOLOGY TO CULTURE
With visions of teamed bioreactors still dancing in the audience’s heads, the remaining panelists brought a mixture of cold water and qualified support to the idea of horizontal automation in their industry. Lars Petersen, head of automation at Roche/Genentech, spoke of his company’s Tech Transfer and Process Platform Initiative, which aims for more robustness in technology transfer in scale-up, the use of standard process platforms, and other new approaches to speeding drug development.
“Horizontal is the way of the future,” Petersen said. “But the culture of how people are thinking is so significant that it could be a bigger issue than technology in the adoption of horizontal automation.” He explained that the sequential manufacturing functions at most companies — laboratory, clinical, and large-scale production — still exist in separate silos and typically resist talking with each other. For instance, the process development (PD) lab may consider compliance issues so important to clinical manufacturing as outside its responsibility and thus reject the inclusion of compliance modules in its own work. Similarly, clinical manufacturing may react negatively when the company introduces SAP systems into its area.
“We want to know the set-up of the process end to end,” Petersen said. “Problem is, the PD lab would typically develop a single step in the process, then hand it over to clinical before it began work on the next step. It all happened incrementally. But we had our PD group develop the process platform, a processdevelopment format, which creates a framework around every molecule.”
Petersen observed that in most companies no single person oversees the development of a product and process from the lab, through clinical, and into operations. Typically, those functions are headed separately by managers who report high up in the organization. “So the process development head is focused on ‘How can I get my process development done first?’ and that is often at the expense of how fast you can get it into clinical. It is an issue that affects the entire industry,” he said. “No one has solved it.”
Ian Allan of Infinity Automation showed one way a process platform can speed process development. “You cannot increase yield, throughput, and so on without truly understanding the process. It’s a simple loop: understand the variation, start to manage the variation, and then build your control limits around it.” But such improvements depend on a continuity of knowledge traditionally lacking in many companies, he said. “Companies lost process understanding when they lost the process engineers who created it.”
Allan elaborated on the technology needed to “understand the variation,” describing a case of real-time monitoring leading to process improvement. Engineers were instructed to use new instrumentation to record pH levels continuously for 50 minutes in a particular phase of the operation. Not only did the method increase operators’ focus — letting them see the data and respond to control rather than depend on alarms — but it also allowed them to conduct “real-time deviation management” and create a template for ideal initial conditions in the selected phase by matching one batch to another.
Francis Sidnam, director of biologics manufacturing & process development IT and global manufacturing & supply IT at Bristol- Myers Squibb, leads his company’s “process robustness” initiative for pharma and biotech manufacturing. He capped the previous presentations with a more detailed look at “process robustness” on the manufacturing side, as exemplified by the new BMS Biologics Drug Substance plant in Devens, MA.
Sidnam gave some background on the plant. Expanding on the company’s Paperless Plant Systems initiative, originally designed to deploy a system of electronic batch records in production, the company is developing a comprehensive “paperless production” model at Devens to be used eventually at all of its plants, including API manufacturing and even non-automated sites. Besides GMP and regulatory compliance, he said the model’s efficiencies will prove just as beneficial in the long run.
DO CEOs & VCs CARE?
During the question and answer period, much of which was technical, I asked the panel in general to describe how the CEOs of life science companies typically view manufacturing issues. Are they normally interested and involved? Do they see manufacturing as an important strategic area, a competitive factor, a resource worthy of investment and optimization?
Responses from the panelists and the audience were naturally cautious. But it was easy to see a consensus that, at least historically and up to recent times, pharma and biopharma top management has not been known for its interest in manufacturing. The C-Level suites in many companies still prefer to keep manufacturing at arm’s length, content to let plant managers run existing facilities along conventional lines in technology, process, and organizational structure. The same was said of how chief executives tend to regard their CMOs — distantly.
If CEOs pay too little attention to manufacturing, how do the primary investors in the most innovative part of the industry — small companies developing new drugs — treat the issue? The BIO Investor Forum helped me see at least part of the answer.
A roundtable of VCs gave a forecast for such companies in the coming year, most believing that the current burst of M&As will continue in 2013 as the total amount of venture capital investment dwindles. It was observed once again that most small companies fail in the transition from Phase 1 to Phase 2 and Phase 3 development, often because they run out of money paying for clinical trials. But they also acknowledged that clinical manufacturing, commonly perceived as a cost factor only to be minimized, can also play a role in development failures.
My take: Too few companies and investors value compound optimization and other supply chain components that can greatly affect drug potency, stability, and delivery — and thus safety and efficacy — in clinical trials. What are some ways all the players — top management, operations, investors, and others — could collaborate to solve a common problem like manufacturing, that sinks so many companies developing potential medical breakthroughs? The life science industry awaits new leadership that can make the critical connection between optimized manufacturing and competitive advantage, as a bridge from industrial to investment strategies.
BMS Creates A Model Plant
At a dinner event during the Emerson Global Exchange, where Bristol-Myers Squibb (BMS) received Emerson’s Innovative Application Award, BMS Executives Chris Stevens and Dave Gleeson spoke on the topic “Manufacturing for the 21st Century” and described the Devens plant in more detail.
The plant produces the biologic drug Orencia (abatacept) for treating rheumatoid arthritis symptoms and slowing joint damage. The facility is recipe-driven and designed to meet ISA88 and ISA95 standards, employing operations management software and a digital automation system for paperless manufacturing. It is integrated with enterprise and plant-level systems, including SAP, LIMS (lab information management system), scheduling and computerized maintenance management systems, process information historians, and off-line instruments.
One specific goal for the company was for all CCPs (critical control parameters) and CQAs (critical quality attributes) to be electronic; automatic control charting was another. BMS also wanted alarm and event notification to comply with Western Electric Rules selected in its hierarchies, generating daily runs to automatically report any WER (Western Electric Rules) violations to the appropriate people, thus avoiding having to search for and report such items individually. The notification system uses the principle of “review by exception,” looking only for variations beyond standard parameters.
Devens is the biologics site, but Francis Sidnam, director of biologics manufacturing and process development, said its model can be applied to pharmaceuticals as well. “In tablet presses, with our dry granulation, we try to have consistent product hierarchies so we can easily deploy them to other sites.” Despite some data connectivity issues, depending on the data source from individual machines in a given site, the company is moving quickly forward with deployment. It is pragmatically adapting off-the-shelf components — hardware and information technology already widely employed in many industries — to existing facilities, rather than building some uniform, proprietary system from scratch.
To make the needed meta-analysis, decisions, and predictions based on all the discrete batch data, the company built the model around its MES system, automatically pulling “contextual data” from its LIMS, ERP, and historian systems. It created “universes” within its product “hierarchies,” which, once developed for the first product in a product type, can be reused for subsequent products of the same type. Building such a model requires starting with clear analytical goals, such as trending, based on batchto- batch comparison, Sidnam said. He emphasized that the system’s data structure must be something an engineer, not just an IT person, can understand. “Process engineers must be familiar with the process and also have the statistical training to look at the hierarchy and understand the data.”
(Thanks to Healthcare Packaging for sharing its reporting on the dinner presentation.)
JARGONICTIONARY
The following is a quick guide to translating the technical terms and acronyms used in this article.
CCPs: Critical Control Parameters
CQA: Critical Quality Attributes
ERP: Enterprise Resource Planning (ERP) systems automate internal and external management information across an entire organization, including finance/accounting, manufacturing, sales and service, customer relationship management, etc. with an integrated software application.
Historian (or Batch Historian): Record of all batches and their associated data.
LIMS: Laboratory Information Management System (LIMS) is a software-based system that supports a laboratory’s operations.
MES: Manufacturing Execution Systems (MES) are information technology systems that manage manufacturing operations in factories.
Product Hierarchy: A catalog of product standards, characteristics, assays, protocols, processes, and other attributes classified and ordered from general (all products in a class) to specific (unique to product or subset).
SAP: An enterprise software application developed by the German company, SAP, used to manage and integrate multiple business processes and functions into one comprehensive system.
Western Electric Rules: In Statistical Process Control, the Western Electric Rules (WER) are decision rules for detecting “out-of-control” or non-random conditions on control charts.