Proceedings of a Workshop
Barriers to Innovations in Pharmaceutical Manufacturing
Proceedings of a Workshop—in Brief
Modernizing pharmaceutical manufacturing is seen as offering a solution to drug shortages and vulnerabilities of drug supply chains, but there are technical, regulatory, and other barriers to doing so. The virtual workshop Barriers to Innovations in Pharmaceutical Manufacturing held on June 2–3, 2020, provided a venue for discussing barriers to innovations in the pharmaceutical industry and included sessions on integration, intensification, and control; innovative processing technologies; and disruptive technologies and convergent innovations. It was hosted by the National Academies of Sciences, Engineering, and Medicine Committee to Identify Innovative Technologies to Advance Pharmaceutical Manufacturing and served as the second information-gathering activity to assist the committee with producing its consensus report.1 This Proceedings of a Workshop—in Brief summarizes the presentations and discussions that took place during the workshop. The workshop videos and presentations are available online.2
REFLECTIONS FROM THE OFFICE OF PHARMACEUTICAL QUALITY
Michael Kopcha, director of the Office of Pharmaceutical Quality in the U.S. Food and Drug Administration (FDA) Center for Drug Evaluation and Research (CDER), opened the workshop by providing his perspectives on barriers to innovation. He began by defining pharmaceutical quality as safety, effectiveness, and freedom from contamination and defects in every dose. He noted several approaches that CDER uses to regulate pharmaceutical quality. Facility inspections are designed and conducted to prevent problems. Quality assessments are conducted on each application before it is approved. Active surveillance and enforcement programs track marketed products and the facilities that produce them. Multidisciplinary intramural and extramural testing and research programs allow FDA to test products on the market and keep apprised of technology developments that might affect product quality. FDA develops policies that support not only its own quality assessments but global harmonization of practices for regulating quality. And outreach and engagement programs communicate directly with the industry and the public to explain quality requirements. Collectively, those approaches give confidence in the quality of medicines that people in the United States use.
Kopcha stated that 62% of drug shortages result from quality issues and that advanced manufacturing could help to address supply disruptions. Advanced manufacturing includes novel manufacturing methods that can improve process robustness and efficiency, novel dosage forms that can improve drug delivery and targeting, and analytical tools that can improve product quality testing, process monitoring, and control. He emphasized that advanced manufacturing is important because it can address the underlying causes of drug shortages, facilitate new clinical modalities, and improve manufacturing efficiency, but he noted important barriers related to quality that he personally thinks must be overcome:
- Regulatory barriers, including the challenges of fitting new technologies into existing regulatory frameworks, practices, and concepts; issues associated with inspecting facilities that have new technologies and with making changes after application approval; and the lack of international harmonization or regulatory convergence.
- Technical barriers, including inflexibility of manufacturing operations, thinking, and management and lack of process analytical technology that is needed to enable real-time quality assurance.
1 For information on the committee’s membership and study task, see https://www.nationalacademies.org/our-work/identifyinginnovative-technologies-to-advance-pharmaceutical-manufacturing.
- Financial barriers, primarily the need for capital to implement a new manufacturing technology and the lack of a clear business case for making a change.
- Logistical challenges, including training requirements for industry, regulators, and investigators; competition for the small pool of needed new talent; issues surrounding intellectual property that make adoption of new manufacturing technology difficult; and lack of strategic partnering in the industry.
Kopcha emphasized the importance of partnerships of industry and regulators to advance manufacturing methods that can potentially solve some of the long-existing problems within the industry. He concluded that it is time to start thinking differently about how pharmaceuticals are manufactured.
DELIVERING ON MANUFACTURING INNOVATION: CHALLENGES AND OPPORTUNITIES
Roger Nosal, vice president and head of global chemistry, manufacturing and controls (CMC) at Pfizer, opened the first session by highlighting substantial regulatory challenges in advancing innovations in pharmaceutical manufacturing. He noted that the pharmaceutical industry has often been characterized as using outdated technology, relying on vulnerable and complex supply chains, having too limited and fixed capacity, and being overregulated. He countered by noting factors that hinder innovations and emphasized that pressure on the industry to accelerate product development actually discourages it from changing its processes and said that both the industry and the regulatory authorities suffer from risk aversion. He said that if an innovation does not improve or provide at least comparable quality assurance related to a product, there is no point in moving forward with it.
Nosal stated that expanding application expectations and standards pose substantial challenges for the industry. A recent analysis at Pfizer showed that the content of CMC applications has increased by 30% over the last 20 years. The question is whether all the added information is necessary. The added cost, time, and effort, for example, needed to conduct extraneous comparability studies often seems to be an exercise in “checking a box” rather than adding value, he said. He continued that the goal is to improve quality assurance, and he suggested several points to consider. First, expanding CMC application content does not increase product quality but only increases the volume of regulatory maintenance and that the emphasis should be on improving quality assurance. Second, product specifications do not control product quality; they confirm it. Third, one size does not fit all; generic products have business profiles different from those of innovative products. He noted that Pfizer abandoned an innovative approach—real-time release testing approved for seven products—because of the maintenance costs and regulatory lifecycle burden; adjustments of probe sensitivity and modifications to model algorithms required postapproval regulatory submissions, which negated the business case for the approach.
The other substantial challenge in the industry is global regulatory divergence, Nosal said. The adoption and implementation of international agreements are inconsistent, the default is prescriptive rather than flexible standards, and regulatory expectations are inconsistent, particularly for models. He added that assessors and inspectors are not trained to see the big picture and that early discussions regarding innovations do not appear to inform the product review process. Many problems arise from inadequate technical training of regulators, adherence to regulatory norms rather than scientific principles, different and conservative perceptions of risk, and confusion between a business risk and a regulatory risk.
To conclude, Nosal offered several suggestions for overcoming the challenges described: (1) refocus regulatory applications on relevant scientific justification rather than regulatory norms, (2) create incentives for the industry to invest in new processes and replace antiquated ones, (3) accommodate innovations that improve quality assurance but allow some adjustments, (4) introduce a new paradigm, such as cloud-based applications, for transparency, (5) leverage and continue to invest in international guidelines, and (6) encourage and expand industry collaboration programs.
Patrick Swann, vice president of the Quality Science and Technology Group at Amgen, highlighted approaches for overcoming some barriers to innovation and reducing product commercialization timelines. He began by describing attribute-based control strategies for complex products. He emphasized that a science- and risk-based strategy should be pursued to ensure patient safety and product efficacy and outlined the steps to establish a patient-centric quality control standard. The first step is to determine attribute criticality, whether attribute testing is required, and, if so, the appropriate testing method. In this step, one conducts (1) a product quality attribute assessment that helps in assessing the risk that an attribute will have a clinical effect and (2) a product quality risk assessment that can link material attributes and process parameters to the critical quality attributes. The second step is to identify the optimal testing scope and control points to ensure that redundant, otherwise unnecessary, or inefficient testing is eliminated. The third step is to use various data sources (a totality of evidence) to establish an attribute range acceptable for maintaining safety and efficacy.
Next, Swann described the need to modernize analytical methods so that attributes of concern can be measured and enable an attribute-based control strategy as opposed to a conventional approach that uses profile-based assays. He emphasized the need to use multiattribute methods that can replace multiple conventional methods and measure specific attributes. Monitoring using multiattribute methods can also support process understanding: one can evaluate product attributes in real time and correlate process parameters with the attributes and thus improve process understanding. Swann stressed that it is imperative to modernize analytical methods to enable next-generation manufacturing that will advance process intensification and analytical integration.
Swann closed by emphasizing the need to use prior knowledge and modeling to overcome barriers and expedite product development. He noted that Amgen has 40 years of experience with commercial and investigational products and provided an example of how stability data could be used to determine the shelf-life of a family of related biologic products. He noted that various criteria would need to be met to justify the transferability of stability data. He continued that prior knowledge could also be used to evaluate and quantify method variability.3 The idea is to use data collected over years to obtain a better understanding of method variability rather than using only a validation exercise that provides the results of an assessment at a single point in time. In summary, he emphasized that the industry needs to focus on relevant attributes, fit-for-purpose technologies, and appropriate knowledge management to advance innovations and accelerate product development.
Christine Moore, global head and executive director of the Global Regulatory Affairs and Clinical Safety CMC–Policy at Merck, discussed the need for regulatory flexibility to support flexible manufacturing. She began by listing several drivers of flexible manufacturing: the movement toward specialized products that are manufactured in smaller volumes, the desire to reduce operating costs, the need to respond rapidly to changes in demand, the need to avoid drug shortages, and the pressure to accelerate drug development. She said that innovations—such as continuous manufacturing, portable manufacturing, real-time analytics, single-use systems, and point-of-care manufacturing—are being pursued to create the flexible manufacturing desired.
Moore stated that regulatory agencies have developed some infrastructure and guidance to support innovations and discussed the status of continuous manufacturing. She defined continuous manufacturing as a system that consists of two or more unit operations into which materials are fed and transformed and from which the processed material is continuously removed. Benefits of continuous manufacturing include shorter processing times, integrated unit operations, a smaller equipment footprint, elimination of traditional scale-up, and a system that enables real-time analytics and control. Technical and regulatory challenges include uncertainties regarding real-time release testing and controls and possible difficulties in tracing materials through the system, predicting disturbances of homogeneity, and segregating potentially nonconforming material. She noted that FDA has approved seven continuous-manufacturing applications, that international regulatory agencies have approved continuous-manufacturing applications, and that regulatory bodies have published guidance and are developing regulatory frameworks and formal guidelines.
Moore highlighted regulatory impediments. Some technologies require lifecycle maintenance, such as updating or recalibrating a model, and regulatory approaches have not advanced to accommodate or enable these technologies. Furthermore, she said that there is no regulatory framework for portable manufacturing and individualized medicine and that many questions will need to be addressed to enable these innovations. She concluded by describing three strategies for overcoming the impediments. First, control-strategy elements need to be appropriate to specific products and process risks, and a science- and risk-based approach should be used. Second, one needs to determine what elements are needed for patient safety and efficacy vs manufacturability and recognize that manufacturing complexity need not equate to regulatory commitments. Third, regulatory bodies need to align dossier contents and regulatory expectations regarding changes after approval; that is, international regulatory convergence is needed. The goal, she said, is the availability of high-quality medicines to patients, and regulatory agencies and the industry need to work together to achieve this goal.
Kelvin Lee, professor of chemical and biomolecular engineering at the University of Delaware and director of the National Institute for Innovation in Manufacturing Biopharmaceuticals (NIIMBL), concluded the first session by discussing the role of public–private partnerships in advancing innovations. Lee stated that NIIMBL is a 150-member organization sponsored by the National Institute of Standards and Technology and focuses on accelerating innovation in biopharmaceutical manufacturing.4 The hope is that collaboration in a consortium will accelerate transformation because no one has to be the first to take on all the risk attached to implementing a new technology—everyone goes first together. Lee stated that the vision is to create flexible, agile manufacturing as noted by previous speakers, and he described NIIMBL’s high-level strategy for process intensification as (1) accelerating adoption of intensified processes already developed, (2) collaborating to develop process technology and equipment for use in commercial production in 6 years, (3) designing a hypothetical facility to determine what technology needs to be developed, and (4) developing technology with platform processes demonstrated with collaborators and in a NIIMBL test bed. To achieve flexible, agile manufacturing, however, he noted that an industry survey conducted by NIIMBL indicated that analytical technologies and technologies associated with upstream operation and process monitoring and control are needed.5
Lee next described an active listening meeting between the industry and FDA to understand the challenges in implementing new technologies in biopharmaceutical manufacturing.5 Companies were asked to consider several questions and were then interviewed individually before the meeting to identify consensus topics and areas of concern. The process revealed that the industry has had a wide variety of experiences with CDER and that the industry can rarely identify a business case for implementing new manufacturing technologies. Lee emphasized that speed to market is a key driver and that delays that might occur with introduction of new technologies pose a business risk that many companies just are not willing to
3 Apostol, I., R. Wu, M. Ko, J-L. Song, L. Li, G. Schlobohm, and W. Szpankowski. 2020. Prediction of Precision for Purity Methods. J. Pharm. Sci. 109(4):1467-1472.
5 Mantle, J.L., and K.H. Lee. 2020. NIIMBL-facilitated active listening meeting between industry and FDA identifies common challenges for adoption of new biopharmaceutical manufacturing technologies. PDA J. Pharm. Sci. Technol. See https://journal.pda.org/content/early/2020/05/28/pdajpst.2019.011049.
take. At the meeting, industry representatives met in the morning to receive feedback on the interview process and then selected several topics—the business case for adoption of new technologies, changes in approved manufacturing processes, interaction between the industry and FDA, and consistency throughout FDA—to discuss in the afternoon session with FDA. The afternoon discussion highlighted the need to address the fragmented regulatory environment, the desire for more informal interactions with FDA, and the industry’s varied experiences with FDA that depend on the experience of the FDA staff.
Lee concluded by offering three policy suggestions. First, a policy should be considered that requires some form of public dissemination of the types of technologies and approaches being evaluated by the FDA Emerging Technology Team. He posited that such a policy might improve submissions by helping to focus discussions around particular approaches and could lead to collaborative activities for demonstrating innovative manufacturing technologies. Second, a policy is needed that provides incentives to support investment in advancing manufacturing technologies; such incentives are few. Third, a policy to create mechanisms for informal interactions between FDA and the industry to discuss and learn about manufacturing innovations would be advantageous. Lee stressed that there is a real opportunity now to move forward together and that that is what these public–private partnerships, such as NIIMBL, are trying to accomplish.
Matthew DeLisa, William L. Lewis Professor in the Robert Frederick Smith School of Chemical and Biomolecular Engineering at Cornell University, moderated a panel discussion with the speakers and the workshop audience. The discussion opened with a question on whether there are criteria for helping a company to decide whether to invest in an innovative technology. Lee noted that there have been heated discussions on this topic but that there appear to be no structured or accepted criteria. Nosal agreed with Lee and commented that one can often see long-term benefits of some technologies that provide an ability to accelerate production quickly but that the cost savings will depend on various factors associated with the specific product. Moore commented that one can see an advantage, for example, of flexible or portable manufacturing for products that might have highly variable demand.
Much discussion centered around incentives that might drive more innovation. Nosal stated that there should be an incentive to focus applications on relevant information and not simply the volume of information. He added that incentives should lead to an environment in which innovations can be introduced more quickly and changes made more easily as the technology dictates. Moore added that incentives are needed to create performance-based regulations that are patient-centric. For example, if a manufacturing process is controlled and well understood and product testing demonstrates attributes acceptable for patients, not all aspects that control the process should be regulatory commitments. As suggested earlier in the session, Swann stressed that the FDA Emerging Technology Team’s publication of its views on technology that it is considering and how the technology could be implemented would constitute an important incentive. Such information could improve the predictability of innovations.
An audience member asked the panelists how a company can accelerate introduction of a new technology when there is little legacy understanding to support its introduction? Nosal responded that joint meetings with regulators from various regions could help by allowing everyone to hear each other’s views and concerns. It is an option that has not been fully explored by the industry. In response to a question on cultivating international convergence and making the regulatory process more effective or efficient, Moore commented that it would be helpful if FDA directly trained regulators in new technology. Nosal agreed but added that everyone needs to look collectively at how international regulatory agencies can align. Lee noted that patient advocacy groups could play a key role in driving convergence and mutual acceptance inasmuch as some products created by innovative technologies are so promising for patients. As a final question, an audience member asked about regulatory risk associated with introduction of a new technology. Nosal noted that if a company is first to introduce a technology, it takes on the burden of educating regulators and handling unexpected issues that can be time-consuming. Moore added that introducing a new technology often requires redundant work to prove that the technology is working as intended, and this means more time and money.
To close the discussion session, DeLisa asked the panelists to provide final remarks. Moore stressed the importance of using a science- or risk-based approach, keeping the focus on patients, and moving toward international regulatory convergence. Nosal added that different incentives will be needed to advance innovations in the manufacture of old products vs new products. Lee emphasized the importance of incentives to accelerate innovative approaches; in the absence of incentives, it will take years to implement what should be relatively straightforward technologies.
BLURRING THE BOUNDARIES: INTEGRATION, INTENSIFICATION, AND CONTROL
Paul Collins, senior director of small molecule design and development at Eli Lilly and Company, opened the afternoon session by suggesting that continuous manufacturing is key to the future of new drug development. He said that the human genome project opened the door to new modalities and that pharmaceutical processing operations need to evolve to produce new drugs. Technical challenges that lie ahead include dealing with undesirable solvents, creating efficient purification and isolation operations, identifying sustainable starting materials, using new chemistry, and scaling delivery vehicles. Unfortunately, he said, pharmaceutical operations today are big, expensive, and designed for multiuse but are confined to 20th century chemistry and small-molecule manufacture. The scale and design hinder production of new modalities and rapid response to global events.
Collins stated that continuous manufacturing might not be the solution to all the problems but is certainly important because it allows processing on a smaller scale and removes or reduces scale-up as the most important process development variable. He emphasized that continuous manufacturing allows one to miniaturize operations to “modules,” so that facility design can focus on moving modules in and out as needed. Multiple groups can design modules that can fit into a facility, and modules can be standardized for common use. Module standardization also eliminates the need to prove that a module will perform equivalently when used in a new setting. He described how a facility can be built by using a modular approach and noted that the difficulties involved in filing such advanced technology need to be addressed. For example, such filings typically result in long discussions with regulators that require addressing many questions and often negative views toward models and process analytical technology given the maintenance requirements. As a solution, Collins suggested that the educational aspect of the filings be divorced from the review. He stressed that sites that use “approved” modules should not have to “prove” new site equivalency and that boundaries and terminologies need to be redefined in light of advances in science and technology. He closed by emphasizing that the drugs of the future and responses to world health crises require the industry to change and that continuous manufacturing provides a framework for addressing such change.
James Thomas, executive vice president and global head of Just Biotherapeutics and president of U.S. operations at Just–Evotec Biologics, discussed the design and construction of next-generation biologic manufacturing facilities. He echoed the goals of earlier speakers to use advanced technologies to produce high-quality biologics with greater speed at a low cost and to create flexible capacity. He described the Just–Evotec Biologics approach to integrated design, which starts with “efficient” discovery, optimizes products, and then uses a highly productive process that leads to flexible, low-cost manufacturing. His company has developed a suite of machine-learning tools that can examine the DNA structure of antibodies and identify where changes can be made to improve stability and thus the potential for commercial success. The key, he said, is to incorporate “developability and manufacturability” into the discovery process.
Thomas agreed with Collins that it is critical to shrink the footprint of unit operations so that a company can optimize its facilities and that one approach is to use continuous manufacturing. He said that using continuous manufacturing successfully for biologics requires a focus on a few critical needs: aseptic conditions throughout the entire production process, process control that can maximize productivity while delivering consistent product quality, and a supply chain for high-quality disposables and consumables used in the process. He described a continuous manufacturing facility in which unit operations are contained in PODs that can be moved in and out and reconfigured. Using that type of facility design allows clinical and commercial processes to operate on the same scale and thus facilitate a seamless transfer of processes. In response to a question, he noted that one advantage of continuous manufacturing for biologics is the elimination of large intermediate storage tanks and the associated plumbing. He concluded that continuous manufacturing in well-designed, relatively inexpensive facilities can deliver flexible capacity for producing the highest-quality biologics at a reasonable cost.
Mauricio Futran, vice president of advanced technology in the Global Tech Services Group of Janssen Supply Chain at Johnson and Johnson, agreed with previous speakers that the future of pharmaceutical manufacturing is a small facility that uses digital and real-time control and is modular and flexible. But he asked, How do we get there? He stressed the value of mechanistic and statistical models and described a case in which his company collected data on a small scale for various processes, created models, and then projected performance at operating scale. When the new plant came online, they were able to run a single confirmatory batch for every product and then proceed straight to validation and commercial production. That approach to technology transfer has provided his company with important benefits, including substantial cost savings, shortened project timelines, and reduced quality-control sampling, project risk, and material use and waste.
To realize the vision of future pharmaceutical manufacturing, Futran emphasized the importance of real-time product quality awareness and the benefits of using advanced monitoring and process control to achieve real-time release of a product. He compared batch processing of tablets with quality-control laboratory testing—in which seven instruments are used to provide the necessary information for release in days, weeks, or a month, depending on the workload of the laboratory—with continuous processing and real-time release in which one instrument provides all the necessary information in minutes or hours. He said that real-time awareness allows examination of why some processes, for example, have better yields or produce particular impurities and that the resulting information allows tighter control and better outcomes and yields. His company anticipates increased use of models and advanced sensors, given the benefits realized so far. He also noted that his company wants to move to modular design as described by Collins and Thomas but that it would be helpful if standards were developed.
Futran concluded by discussing challenges faced by the industry. He said that there are internal challenges because the incentive is to minimize questions and the timeline for approval and that new technologies typically lead to more questions and concerns about longer approval time. New technologies also require resource investment and a culture change to value a data-rich environment. He stated that external or regulatory challenges loom large because the regulatory world is based on offline testing of batch processes, which do not translate well to new manufacturing technologies that allow varied batch sizes, inline analytics, and higher-fidelity methods for detecting batch to batch variation that potentially provide a superior approach to control. He stressed that the best practice for innovation involves interacting with FDA so that one can discuss how to interpret old regulations and ways to bridge perspectives and interpretations. The greatest opportunity, he noted in closing, is to transform the regulatory framework so that companies can move from procedural control to engineering control; moving from sample and test to real-time awareness constitutes a fundamental shift for ensuring control and reliability but one that is essential for advancing pharmaceutical manufacturing.
Kerry Love, chief executive officer and president of Sunflower Therapeutics, speaking from the perspective of a contract development and manufacturing organization, described the development of right-sized automated manufacturing systems for protein production. Her company has created two systems: DAISYTM, which provides laboratory-scale production (0.1–1.0 kg annual capacity) to support early discovery research, and DAHLIATM, which provides pilot-scale production (1.0-10 kg annual capacity) to support clinical development. She noted that seamless process transfer can occur between the systems and listed several critical technologies that enabled system development. First, the microbial host, Pichia pastoris, that is used in the systems does not get contaminated by viruses, has relatively few secreted host cell proteins, and thus offers dramatically simplified purification operations. Second, biology-driven expression engineering allows molecular design to enhance quality and manufacturability. Third, an end-to-end approach to process development minimizes interfaces and provides data for understanding the link between processes and specific products that can aid rapid process development for new products.
Sunflower systems incorporate aseptic single-use components, continuous operations, process automation, and functionally closed processing, Love stated, and this holistic design should lead to reduced testing of the end product, which for Sunflower is the bulk protein drug substance. She noted further that three conventional unit operations—upstream production, downstream purification, and formulation—have been combined into one with no user intervention and that biological variance has been reduced by using simple hosts that can be easily engineered. She continued, saying that Sunflower’s standardized systems provide the opportunity to accumulate many process performance datasets, and this should lead to the use of process control data instead of testing to ensure product quality and to satisfy release requirements. She noted that another benefit of the systems is that they support flexible capacity. Such systems could be deployed globally to create a network of manufacturing centers that would be capable of operating the same process that was developed in a single location. She concluded by saying that these systems would dramatically streamline technology transfers between process development and manufacturing and that such an agile global network would enable preparedness for future pandemics and improve global access to biologics.
Kim Wolfram, director of regulatory CMC at Biogen, discussed the alchemy of process control, which she defined as the transformation of designated process controls into a control strategy that is derived from prior knowledge and product and process design and that ensures process performance, product quality, and sustainable supply. She added that regulatory flexibility is needed and that the strategy needs to be built on a foundation of trust, diversity and inclusion, and patient-centricity. She emphasized the need particularly for diversity and inclusion and stated that input from a wide variety of colleagues and external partners is needed to advance pharmaceutical manufacturing and integrate the entire drug production process.
Wolfram presented her views on the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) Q12 guidelines that are meant to facilitate management of postapproval CMC changes and thus enable advanced process control strategies. She hopes that the guidelines increase the use of science- and risk-based approaches and lead to regulatory convergence. She listed several advances that could increase product supply robustness and noted that because of the regulatory complexity and the potentially long global approval timelines, some companies are not pursuing advances, but she countered that one needs to consider the risk of not implementing changes. She stressed that advanced analytics and modeling need to become parts of the culture to improve understanding of products, process, and design and that developing a common or standardized language for the tools will be essential.6 She concluded that although it is important to consider the viability and feasibility of new technologies, one must not forget the human context—who the patients are and what their needs are—that needs to be embedded into product profiles.
Jason Starkey, senior director of analytical research and development at Pfizer, continued the discussion of control strategy as a critical aspect of manufacturing innovation. He agreed with Moore that there is no “one-size-fits-all” control strategy and said that development of Pfizer’s control strategy starts with the molecular design. It is critical, he said, to understand the mechanism of action, potential for immunogenicity, and any adverse feature that needs to be eliminated. Next, one must consider the expression system and manufacturing controls and then apply purification controls to remove host cells and impurities. Identifying the best approach to characterize product quality is paramount. The final considerations are drug formulation, packaging, and delivery device if applicable.
In developing the control strategy, Starkey said, a substantial hurdle is defining the product quality attributes. He noted that there are tools for assessing structure–function relationships and product degradation, stability, and purity, but better tools are needed, for example, to assess immunogenicity potential. He emphasized that advanced process control strategies allow one to ensure product quality by understanding and monitoring process and analytical controls, not simply by product testing. The key, he said, is to understand how process controls affect product characteristics, such as glycosylation on antibodies, which ultimately might affect potency. He highlighted the multiattribute method that can analyze multiple product quality attributes simultaneously, provide much richer information, and allow automated monitoring and quantitation of attributes. Starkey stated that several companies are exploring this method for process control and that the collective effort is helping to reduce doubt and uncertainty and ultimately to advance its acceptance. He did caution that in the end all the testing needs to align.
6 Romero-Torres, S., K. Wolfram, J. Armando, S.K. Ahmed, J. Ren, C. Shi, D. Hill, and R. Guenard. 2018. Biopharmaceutical Process Model Evolution—Enabling Process Knowledge Continuum from an Advanced Process Control Perspective. Am. Pharm. Rev. 21(4):62-71.
Starkey closed by discussing how process control strategies will need to be modified, given new manufacturing processes that involve continuous operations. Pfizer is developing a continuous upstream process with periodic downstream purification for some products, and the company is trying to understand how setting process variables in the continuous operation will affect product quality attributes by using such tools as the multiattribute method. Given the innovations in manufacturing and control strategies, Starkey agreed with Wolfram that ICH Q12 could be transformational in that it could harmonize change management, facilitate risk-based regulatory oversight, support continual improvement, and enable strategic management of postapproval changes. He hopes that the transformation will improve transparency and compliance, ensure supply reliability, reduce postapproval submissions for extraneous manufacturing changes, optimize processes and reduce product variability, standardize expectations for postapproval validation, and optimize resource deployment for assessment and inspection. Starkey concluded that introducing innovative technology remains a challenge in the global setting; when a company that develops regulatory submissions intended for a global supply chain faces regulatory feedback or standards that are out of broader international alignment, that discordance encourages a company to follow the path of least resistance for global registrations.
Stephen Hadley, senior program officer for vaccine development at the Bill & Melinda Gates Foundation, moderated a panel discussion with the speakers and the workshop audience and began by asking the speakers to reflect on the challenges raised during the session. Thomas noted that the industry tends to be conservative and that there is a risk associated with adoption of a new technology, especially when one already has a process that works. He added that it requires a capital investment, which small companies are often not able to make. Starkey continued that the technical challenges should not be dismissed but agreed with Thomas that there are risks and that there will need to be a period in which technologies are running in parallel to “de-risk” the new technology. Collins added that the power of the incumbent is a big problem—that is, why change a process that works?—and that new technologies have a greater chance of being adopted when one is dealing with new modalities, which will require new approaches. Futran noted, however, that starting with legacy products is a great way to demonstrate and prove a technology; once the technology is proven, it can be applied to other platforms.
Next, the speakers were asked several questions related to their use of data science and modeling. Futran commented that he typically does not have big datasets to allow deep learning but that his company has worked to develop and apply the digital twin technology. When questioned about the timeline for inclusion in filings, Futran noted that there is still some work to do but that FDA should see this technology in its filings within 5–10 years. Thomas and Love noted that they do work with big datasets and that their hope is that interrogation of these datasets leads to optimization of processes and control strategies and ultimately improved product quality. In response to a question about modeling needed to connect modular units, Thomas commented that his company uses process economic modeling that considers all operations in a facility to optimize processes and reveal opportunities for improvement. Collins said that to have control in a plug-and-play system, one needs some fundamental first-principle models to describe unit operations and programs that can communicate with each other through the distributed control system. Starkey added that there is some recognition of that need in the revision of the ICH Q2 guideline to include use of models and parametric data; the revision will hopefully facilitate advanced control strategies. Responding to a query about data-science capabilities in their companies, Collins, Futran, and Starkey stated that they have been working to increase data-science competence at either a department or company level.
The session closed with a discussion of the potential for alternative hosts in the production of biologics. A workshop participant noted that alternative hosts, such as P. pastoris, have been investigated unsuccessfully in the past. Love emphasized that there is a resurgence in interest in alternative hosts, given the emphasis on accelerating drug production. She added that science has entered an unprecedented era in which simple organisms can be engineered to do complex post-translational modifications—that is, they can be modified to become protein factories. The industry is looking for fast, low-cost, high-quality solutions and looking to alternative hosts to provide them, she concluded.
INNOVATIVE PROCESSING TECHNOLOGIES
Susan Hershenson, deputy director of CMC, and Ping Zhao, senior program officer of quantitative sciences at the Bill & Melinda Gates Foundation, introduced the session on innovative processing technologies by sharing their perspectives on drug development for global health. Hershenson began by stating that different product formats are needed to address patient needs and gaps in global health systems. Her organization focuses on low-income, low-resource countries that have a shortage of health-care providers, weak supply chains, challenging transportation infrastructures, and populations that have difficulties in accessing health-care facilities. Those challenges put a premium on products that are easy to use and administer, that minimize the need for patients to have frequent contact with health-care providers and facilities, that have enhanced stability, and that are light, compact, and easily transportable. She said that a primary goal of her organization is to increase availability of safe, effective, and modern contraceptives to millions of women who lack access. Formats that are used in high-income countries often do not meet the needs of women in many countries. Some of the innovative options that her organization is funding are long-acting injectables, biodegradable implants, programmable implants, and oral contraceptives that offer long-acting, well-controlled delivery.
Hershenson noted the common misconception that global health systems can rely on old products, but the unmet needs require innovations, especially manufacturing innovations to enable a variety of dosage forms. She said that product development and manufacturing challenges include sophisticated technologies required to produce complex dosage forms and the need for new devices to deliver products, aseptic manufacturing conditions, long shelf-life, and cost-effective approaches. Zhao elaborated on the regulatory challenges. Given that all product candidates make use of well-characterized contraceptive hormones that have established track records and safe and effective long-term administration, Zhao suggested that the regulatory pathway could be accelerated. He presented a streamlined program that uses a model-informed, exposure-based paradigm to eliminate phase II and potentially phase III clinical trials. He noted that FDA and his organization have entered into a collaboration to assess the feasibility of this innovative approval pathway and emphasized that facilitating knowledge-sharing and collaboration will be the key to acceptance of a new regulatory approach.
Brent Lieffers, senior director of operations at Singota Solutions, discussed modern aseptic processing and challenges associated with validation. He stated that the trajectory of drug development for biologics involves delivery methods that are parenteral and bypass the gastrointestinal tract, so aseptic processing is required. He emphasized that the goal is to prevent or eliminate contamination of the product with microorganisms or particles and that the risk of contamination is greatest when the product is being transferred to the primary container, a container that will ultimately deliver the dose to the patient. The biggest sources of contamination are people, so the solution is to remove operators from the process and the environment. That can be accomplished by using a robust automated process that separates the operator from the environment via a gloveless, robotic isolator. Such units allow effective decontamination, eliminate container–container contact, remove the need for human intervention, and minimize the amount of process-specific tooling. The gloveless aspect is important because it removes the operator from the process, eliminates a source of isolator integrity failure, and shortens decontamination aeration times.
Lieffers stated that there are barriers to implementation of this innovative technology. More regulatory guidance documents that address isolator use are needed, although there are some excellent industry evaluations and publications. He stressed that the barriers are the deeply established validation approaches associated with aseptic processing that are not particularly applicable to the new technology. For example, he said, it is not necessary with this technology to mandate a media fill trial to estimate the contamination risk posed by every combination of container type, closure, fill volume, and batch size, and it is not necessary to require extensive environmental monitoring when operators are no longer part of the process. He closed by encouraging regulatory agencies to use a risk- or science-based approach as suggested by Moore and emphasized the importance of education for regulators and the industry.
Patricia Seymour, a managing director at BDO, discussed the need for novel formulations to improve biopharmaceutical stability. She first noted the many challenges in manufacturing large biopharmaceuticals, but she said, biopharmaceuticals offer many advantages, primarily the promise of success in treating an array of illnesses and diseases. She focused her discussion on monoclonal antibodies (mAbs), enumerated positive attributes of mAbs as drug candidates, countered with some of the technical challenges in their manufacture, and highlighted two issues—degradation and immunogenicity—that are particularly problematic. She described various degradation pathways and noted that developing formulations to prevent degradation can be complicated. Immunogenicity concerns, she said, have led to increased regulatory focus on control of product-related impurities, and developers and regulators are increasingly focused on formulations that would limit the formation of these impurities. Because of the concerns about degradation products and impurities, she said, there is a growing need for excipients that are fit for purpose in meeting the complex challenges of formulating, manufacturing, and delivering new products.
The problem, Seymour said, is that formulators typically use only buffers and excipients that are included in the FDA Inactive Ingredient Database or that comply with standards of the International Pharmaceutical Excipients Council; this approach provides few options for manufacturers. Given the few options, FDA is interested in establishing a pilot program to review toxicology studies that is independent of the various application processes to evaluate the safety of novel excipients. To identify potential novel excipients that might have a public health benefit, FDA asked the industry how drug development challenges could be addressed by using these novel excipients. Industry responded by highlighting various agents that could improve solubility and stability, reduce viscosity, prevent aggregation, enable product delivery, and facilitate fabrication of novel systems. Seymour described several recently developed novel excipients that show promise for extending shelf-life, reducing viscosity, and reducing the potential for immunogenicity. However, she concluded that incentives, such as market exclusivity for the development of novel excipients, will probably be needed, and she emphasized that the FDA pilot program could usher in a renaissance in formulation development of biopharmaceutical products, especially if FDA makes data available and publishes its reviews.
Mansoor Khan, a professor and vice dean at Texas A&M University, described practical aspects of the development-to-approval pathway for innovative formulations and delivery technologies. He first mentioned drugs and delivery technologies that are well understood but are lagging in submissions and approvals and then focused on ones that are well understood and approved but might have barriers to effective marketing and use. He noted that there can be a wealth of information but that it often takes guidance from FDA or other regulatory bodies to move the development and approval process forward. Even with guidance, he said, it can take years for industry working with regulatory agencies to understand and implement the guidance for specific technologies. In addition, there is often a disconnect between agency staff who engage in initial discussion on technology and those who review the applications, and this disconnect emphasizes the need for staff training.
He said that innovative technologies, such as 3-dimensional (3D) printing, can also pose some issues inasmuch as defects in a new product might not be similar to the defects in traditional products and that there might even be questions about how to categorize the product created. Khan concluded that the keys to developing and implementing innovative technologies are to continue to develop strong science-based policies, to use internal and external resources to promote and publicize innovative products and manufacturing, to understand past recalls and connect solutions with modernization of pharmaceutical manufacturing, and to recognize and promote advances in science in the regulatory agencies.
To provide insight into the barriers that a company can face, Jae Yoo, chief technology officer at Aprecia Pharmaceuticals LLC, described his company’s journey from technology development to FDA approval of 3D printing of pharmaceutical dosage forms. He noted that there are various methods of 3D printing and that his company uses the powder–liquid deposition (binder jetting) approach. He stated that a substantial technical barrier in developing its technology was achieving speed and scale for a successful commercial operation. The company was able to do that by creating a process that used a continuous loop with stationary powder spreaders and printheads, constantly re-examining each step, and systematically refining the technology. Another technical challenge was to develop a quick, automated, repeatable process to separate the dosage forms from the unused powder, which is recirculated into the system. Production and testing of many different batches were also required to gain process understanding. A benefit of the testing was that it led to the discovery of many opportunities to reduce process variability and improve product quality.
In addition to technical challenges, there were regulatory challenges, Yoo said. Unbound powder is captured and fed back into the system, and how the recirculation might affect product quality had to be determined. The company used a data-driven approach, which required extensive testing along various points in the process and on many attributes, to generate process understanding, demonstrate product consistency, and provide confidence. The most difficult regulatory barrier in using innovative technology might be managing the process of continuous improvement, Yoo said. Each change requires a substantial effort to prove that the process still works and to validate and launch the product from the new platform. He suggested that companies need an incentive to continue to innovate, given the constraints. He concluded with a few reflections: barriers should be expected in connection with innovative technologies, finding the right problem is often harder than solving it, perceived barriers might be more difficult to overcome than real ones, continuous improvement requires careful planning and execution, and balancing product vs production capacity is crucial for continued innovation.
Steven Nail, principal scientist in research and development at Baxter Pharmaceutical Solutions, closed the session by discussing barriers to improving process efficiency in pharmaceutical freeze-drying. He first described the traditional freeze-drying process and noted that its advantage is that it enables removal of water at low temperatures from thermally labile materials and thus avoids the heat associated with more traditional drying methods. The disadvantage, he said, is that the process is inefficient with cycle times measured in days. He said that there are also uncertainties in the traditional approach: the “edges” of product- and equipment-related failures are not known, and manufacturers generally do not know the optimal processing conditions.
Nail listed several barriers associated with the traditional approach. One is the failure of not thinking long term. Early process development is typically handled by research and development groups, so the priority is simply to produce sufficient product for clinical trials. Process efficiency is a secondary concern, and cycle optimization is often not done even when the product is approved and manufactured for broad use. He emphasized that that barrier could be overcome by constructing a map of all process conditions that produce an acceptable product; such a map would indicate the edges of failure, and he described how it could be generated. A second barrier, he said, is the regulatory authorities’ expectation that companies will follow a design-of-experiments approach to establish process conditions. Nail stated that his company does not follow that approach but prefers a method that uses first principles. The goal is to have freedom to operate around previously determined optimal process conditions. A third barrier is the physics, he said. To make freeze-drying efficient, the process needs to be redesigned, and alternatives to traditional freeze-drying in vials need to be pursued. He highlighted two alternative approaches—spray freeze-drying and continuous freeze-drying—that are worthy of attention and that show great promise for improving the process and shortening drying time from days to hours. He concluded that process improvement is needed, given the increase in large-molecule pharmaceuticals and the fact that about half of them are freeze-dried.
Kelley Rogers, NIIMBL technical program manager in the Office of Advanced Manufacturing at the National Institute of Standards and Technology, moderated the panel discussion with speakers and workshop audience. Several questions were raised about the 3D printing process for pharmaceuticals. In response to a question on the effect of recirculation on product quality, Yoo acknowledged that material can be recirculated several times and that his company specifically investigated that phenomenon and showed that product quality was not affected even with a high degree of recirculation through the system. He emphasized that the experiments that his company conducted to gain a complete understanding of the process and the data generated helped in talking with regulators in the approval process. Yoo was asked about process improvements that were not pursued by his company for its first product approval. He responded that his company wanted to focus the regulatory discussions on the 3D printing process rather than other innovations, such as continuous blending and tabulating or process analytical technologies. He acknowledged that his company feared that regulatory barriers might arise if “too much” innovation were in a single filing.
Rogers asked the speakers to speculate on technology that FDA might see in 5 years and on disruptive technology that could appear in 10 years. Lieffers said that innovations associated with small-batch production for personalized medicine and innovations associated with testing the sterility of parenteral medications are on the horizon. Seymour stated that there will be new excipients in the 5-year period and innovations that eliminate the need for refrigerating products in the 10-year period. Nail suggested that upgrading control systems to optimize processes should happen in the near term. In 5–10 years, he said, FDA should see innovations in freeze-drying technology, such as controlled nucleation, spray freeze-drying, and continuous freeze-drying. Khan hoped that there would be implementation of technology that can monitor and ensure product quality in overseas manufacturing. Yoo said that there will be innovations in 3D printing; his company is working to make the process better, faster, cheaper, and more accessible or mobile. He speculated that in 10 years there could even be 4D printing. Hershenson hoped that regulatory approaches would evolve soon that facilitate approval of dosage forms that use an active pharmaceutical ingredient that has already proved to be safe and effective. In 10 years, she hopes, mobile manufacturing units described earlier in the workshop will become a reality because they would greatly strengthen global supply chains.
To close the session, the panel was asked whether industry groups or partnerships are moving with the right speed to identify and communicate best practices for new technologies or whether something else needs to be done. Lieffers commented that the industry group to which his company belongs has shared innovative methods, and this has helped to develop a unified message to communicate to FDA. He noted how helpful the FDA Emerging Technology Team has been in providing feedback to his company as it develops innovative approaches. Seymour emphasized that there are not universal “best practices” and that regulators are willing to accept a spectrum of information. She said that the industry and regulators need to consider who defines best practices and how they get applied to all the various companies and innovators. She noted that big companies that can afford to generate hundreds of pages of data and documentation for submission often set the bar but that one should consider what is needed from a regulatory perspective. Hershenson hoped that FDA might consider collaborative reviews or sharing dossiers with other regulatory agencies; there are best practices that could be shared and whose sharing would help an industry that is truly global.
DISRUPTIVE TECHNOLOGIES AND CONVERGENT INNOVATIONS
Gustavo Mahler, managing partner at Dynamk Capital, began the final session by offering a perspective from an investor in start-up life science industrials, which he described as companies that drive innovation by developing tools and services that increase yields and productivity and reduce costs of discovery, development, and manufacturing of biopharmaceuticals. He provided his view on innovations in biopharmaceutical manufacturing that could appear in FDA submissions in the next 5–10 years. He said that start-ups are innovating in cell-line development and are focused on high-yield systems that use alternative hosts, synthetic biology, and high-throughput selection of high-producing cells. Companies are also working on cell-free systems. Single-use bioreactors are an important innovation that has recently emerged, and companies are refining this technology by incorporating process intensification methods. Given the increase in titers upstream, he said, downstream processing is an area that has major challenges and that companies are developing alternate cell separation methods based on physical principles, continuous chromatography, membrane-based chromatography, and single-use concentration equipment to address the challenges. Regarding the final production stages, Mahler noted that alternate formulation and active-pharmaceutical-ingredient fill and storage methods are being developed to achieve high concentrations, temperature stability, and better protection of the ingredients after formulation. As a final area of innovation, he noted the creation of new software applications to design and control processes better and the development of multiple options for in-process control technologies and high-throughput analytical technologies, such as inline or offline metabolite monitoring or analysis, fast-separation methods to replace traditional methods, and cell-based in vivo analysis using microfluidic devices. He concluded that funding partners that understand market dynamics and commercialization of new products are crucial for accelerating innovation in bioprocessing but that it is important to remember that new products take several years to enter the market and will probably be adopted first for clinical development of biopharmaceuticals.
Amy Jenkins, program manager at the Defense Advanced Research Projects Agency (DARPA), discussed novel manufacturing approaches to enable rapid responses. She noted that DARPA invests in technologies that are in early developmental stages and will not be ready for commercial applications for many years, possibly decades. One area in which DARPA has great interest is the manufacture of vaccines within days of sequencing the genome of a pathogen. The goal is to have a vaccine approved for use within weeks to confront disease outbreaks and pandemics. Her group is specifically focused on manufacture of nucleic acids. Because nucleic acids might be used as therapeutics as opposed to vaccines, it is not clear where these products will be regulated, and thus this topic should be of interest to the committee that is hosting this workshop. At first, the focus of her group was on finding antibodies, creating nucleic acid constructs, and delivering them to patients, but the group soon recognized that manufacture of high-quality nucleic acids was going to be a bottleneck, given all the challenges associated with cell systems. So, it recently launched a program whose goal is to develop a cell-free system for production of nucleic acids. She noted that challenges for the upstream process will be polymer length, the need for an error-free synthesis, and ensuring simplified starting materials, and challenges for the downstream process will be creating automated production and integrated quality control and product identity assessment. The hope, she said, is that one day there will be an end-to-end system that can manufacture high-quality nucleic acids that can eventually be miniaturized so
that the technology is deployable. In response to a question, she noted that the technology envisioned will not be scalable to produce large quantities but that the technology itself can be deployed to many locations. She acknowledged that the regulatory hurdles could entail trying to approve a novel technology and a novel product, such as a linear-DNA vaccine or therapeutic.
Noubar Afeyan, chief executive officer and founder of Flagship Pioneering, discussed innovations in life-science platforms and some challenges. He began by describing three novel drug modalities that modulate the gut microbiome. The first encapsulates a consortium of bacteria that is designed to restructure the microbiome and modulate disease pathways and is formulated for oral delivery in accordance with good manufacturing practices (GMPs). He listed several manufacturing and quality-control issues that will need to be solved and emphasized the challenges of producing anaerobic microorganisms in spore form and establishing the identity, purity, potency, and safety of a product that is a mixture of microorganisms. The second novel drug modality involves monoclonal microbials that are isolated, fermented, and purified in ways that are similar to those for the manufacture of other pharmaceuticals, although advances beyond current practices were required to produce a specific strain in large quantities and formulate it for oral delivery. The third example was the manufacture of glycans (complex oligosaccharides) to modulate the metabolic profile of the microbiome. For the glycans, a manufacturing process similar to that for small molecules was developed. The process produces many diverse glycans by using small-batch synthesis, integrates multiple analytical methods for structural characterization, and is scalable and transferable.
Afeyan next highlighted two innovations in novel “delivery” vehicles. He first described a new platform that uses anelloviruses for gene delivery. Technical challenges include cell and viral genome engineering, efficient capsid assembly, process design and scale up, supply-chain optimization, and product design for specific tissue targets. The second example highlighted the development of customized exosomes as targeted delivery systems for proteins, RNA, or DNA. He noted that successful industrialization of exosome manufacturing required development of proprietary centrifuge-free purification and proprietary analytical methods to confirm product quality, potency, and consistency. He added that intensified continuous manufacturing is being investigated as the next-generation platform to improve productivity and efficiency and reduce the equipment footprint.
Afeyan concluded his presentation by describing three innovations in cellular therapy. The first involves the use of a novel biocompatible matrix to protect engineered human cells from immune attack and fibrosis. The innovative and scalable automated encapsulation system that was pioneered provides both protection and durability that are commonly lacking in the delivery of cellular therapies. The second innovation involves externally primed T cells with slow-release potent immune agonists to target tumors. The manufacturing process is both scalable and cost-efficient. The third innovation involves the engineering of enucleated red blood cells for use in several therapeutic categories. Manufacturing challenges include substantially improving volumetric productivity, lowering operating costs, and securing raw materials that can ensure safety and supply continuity for the pipeline. In closing, he said that each innovation described presents new challenges but that all indicate how far biological therapeutics have come over the last 3 decades.
Todd Przybycien, a professor in the Howard P. Isermann Department of Chemical and Biological Engineering at Rensselaer Polytechnic Institute, moderated a panel discussion with the speakers and the workshop audience. Most of the discussion centered on various aspects of advancing innovations. The most important criteria for success, Mahler said, are that the technology can meet GMPs, that the supply chain is reliable, that the developer is financially sound, and that the data show that the technology works. He added that including people who have quality-assurance experience early in the development process is critical. To advance novel products, such as microbiome therapeutics, Afeyan noted the importance of working closely with FDA to identify and address issues associated with regulation. Mahler agreed and added that a challenge is the validation procedure; a company will need to establish clear criteria and provide a good rationale for the validation approach for novel systems. The one who makes the first submission will suffer the pain of discovering what it takes to validate the system, he said. In response to a question on how to assess viable investment opportunities, Mahler said that technologies that solve problems or enable someone to take a product to market are clear choices and that technologies that could provide substantial advantages in reducing costs or complexity are also attractive. Przybycien asked the panelists whether a database of common regulatory pitfalls for various types of innovations would be helpful. Mahler and Jenkins agreed that such a database would be helpful for driving innovation and emphasized the importance of clear FDA guidance on validating new technologies. As a final follow-on question, the panelists were asked whether there is a “playbook” that provides regulatory guidance to innovators. Afeyan was not convinced that such a playbook would be useful for innovative technologies and cautioned that knowing what worked or did not work in the past might not apply today, especially with regard to biologics. Mahler countered, however, that there are several basic rules to consider: understand the regulatory implications of the technology being developed, consider how the product will be commercialized and its quality ensured, understand the supply chain and determine who will be the suppliers, and do not underestimate the investment costs in either time or money.
DISCLAIMER: This Proceedings of a Workshop—in Brief was prepared by Ellen Mantus as a factual summary of what occurred at the workshop; no committee member had any role in drafting or reviewing this proceedings. The statements recorded here are those of the individual workshop participants and do not necessarily represent the views of all participants, the committee, or the National Academies.
REVIEWERS: To ensure that it meets institutional standards of quality and objectivity, this Proceedings of a Workshop—in Brief was reviewed in draft form by Christine Moore, Merck & Co.; Steven L. Nail, Baxter Pharmaceutical Solutions; and Kathryn Stein, Kathryn Stein Consulting. The review comments and draft manuscript remain confidential to protect the integrity of the process.
This activity was supported by the U.S. Food and Drug Administration under grant 10004526. Any opinions, findings, conclusions, or recommendations expressed in this publication do not necessarily reflect the views of any organization or agency that provided support for the project.
Suggested citation: National Academies of Sciences, Engineering, and Medicine. 2020. Barriers to Innovations in Pharmaceutical Manufacturing: Proceedings of a Workshop—in Brief. Washington, DC: The National Academies Press. https://doi.org/10.17226/25907.
Division on Earth and Life Studies
Copyright 2020 by the National Academy of Sciences. All rights reserved.