Research and development is not completed until whatever has been developed can be successfully manufactured and deployed, Yamada said. Manufacturing of products for emergency responses can be complex because the products are often for problems that have not yet occurred. There are decisions to be made regarding how much to produce, and at what cost. The cost of manufacturing vaccines is quite high, and is very much dependent on volume (i.e., the more doses made, the lower the cost per dose). Decisions such as whether to use adjuvant or not in global pandemic influenza vaccines can significantly affect manufacturing capacity. Issues on the ground can affect supply chain or deployment, Yamada said. Another critical issue is who receives an intervention first in an emergency situation, especially when it takes time to produce. Yamada noted that, in the case of pandemic influenza, those who got the vaccines first were those who could pay for them, not necessarily those who needed them most. When vaccine manufacturing becomes nationalized, as happened with pandemic influenza vaccine, the output from each country’s manufacturing facility serves that nation first, before any doses are exported. This creates difficult moral and ethical dilemmas, should the global need be greater than the national need. Participants discussed needs and gaps in current manufacturing practices as they apply to the production of medical products for use in public health emergencies, and approaches for delivery and deployment of products that are manufactured outside of the affected region. (Highlights and main points are summarized in the box below.)
Highlights and Main Points Made by Individual Speakers and Participantsa
- Research on manufacturing processes could fill significant and urgent gaps in knowledge about how to supply needed products. For drugs, a switch from batch to continuous manufacturing approaches could benefit both routine and emergency production. For multipurpose solutions such as platform technologies, it is possible that process development, validation of the scaled-up commercial process, and regulatory approval of the platform could be completed in advance of a crisis, potentially facilitating rapid “plug-and-play” manufacturing. (Hamburg, Venkayya, Yamada)
- Participants discussed the logistical and financial challenges of maintaining spare or idle manufacturing capacity for emergency use. One approach is to establish a network of manufacturers (e.g., through partnerships with companies, contract manufacturing organizations, and publicly funded infrastructure) that could rapidly scale up production of products. (Ella, Marks, Venkayya, Yamada)
- Stockpiling is not restricted to final, filled product; product can be stockpiled at an intermediate manufacturing step for finishing when needed (e.g., bulk vaccine stocks), potentially saving space and extending expiration. (Ella, Hall)
- Those at greatest risk, and with most need, are often those with the least access to care and interventions (financially and geographically). Equitable access frameworks that are defined before the next crisis could accompany and inform medical product development. Nationalizing stockpiles can create unnecessary scarcity of products for those most in need. (Hall, Ripin, Venkayya, Yamada)
- Better demand forecasting could create much-needed improvements. Participants discussed the difference between need (number of cases) and demand (what the health system actually has the capacity to deliver), and the potential of scenario-based planning. (Hall, Venkayya, Yamada)
- The local community is not only a beneficiary but also a partner in the delivery of care. Deployment of products can include community care workers, lay providers, and untrained health workers. (Ripin)
- A broad base of support could reinforce the supply chain spectrum, from international air freight service to local trucking services. Also discussed was the critical role of the international military in deployment of facilities, services, and capabilities in a crisis, and local armed forces in supporting logistical and operational functions. (Hall, Pauwels, Yamada)
a This list is the rapporteurs’ summary of the main points made by individual speakers and participants and does not reflect any consensus among workshop participants.
Some of the most significant global manufacturing capacity building and stockpiling efforts in the past several decades began with the preparations for pandemic influenza in the early 2000s, following the emergence of H5N1 influenza as a potential pandemic threat. Venkayya said that political leadership of countries around the globe recognized the potential for a health catastrophe to set off a cascade of events affecting every element of society. He referred participants to the U.S. National Strategy for Pandemic Influenza and the associated Implementation Plan.1 The strategy lays out the key principles around preparedness, response, communication, and other planning for a pandemic, including engagement of all segments of society. The implementation plan, which Venkayya noted is still being followed, outlines over 300 actions for the departments and agencies across the U.S. government that would have a role in a pandemic response and includes guidance for state and local authorities, and for individuals and their families. At that time, he said, the concern was that existing vaccine production capabilities would not be able to keep up to meet the demands of a second wave of infection, as was seen in the 1918 pandemic.
Manufacturing and stockpiling targets for 2006 outlined in the U.S. strategy and plan included domestic vaccine production capability to supply enough doses for the entire American population within 6 months of the declaration of a pandemic, stockpiling prepandemic vaccine to immunize 20 million people, and antiviral drugs to treat 75 million people, with a domestic stockpile of 6 million doses for containment of an outbreak. A $7 billion budget was put forward, half of which was dedicated to enhancing vaccine capacity and developing new vaccine technologies (De Gregorio and Rappuoli, 2014).
1 See http://www.flu.gov/planning-preparedness/federal/index.html for these and related documents (accessed October 30, 2015).
The pandemic influenza model is not necessarily what is needed for emerging infectious diseases, Venkayya said, but it is a good starting point. For pandemic influenza preparedness, the changes in the ecology of the viruses in both the human and animal reservoirs are well understood, and there is massive global vaccine production capacity for seasonal influenza that could theoretically be repurposed toward pandemic vaccine production. This is not the case, however, for emerging infectious diseases. The challenge is how to deal with the “unknown unknowns.”
Venkayya referred participants to the Public Health Emergency Medical Countermeasures Enterprise (PHEMCE) strategy and implementation plan,2 which details what the U.S. government is doing to develop and stockpile agents to address both known threats for which it is unknown if, when, or where they will become a pandemic (e.g., Middle East respiratory syndrome [MERS]) and completely unknown threats. Experience with pandemic influenza and other outbreaks brought to light some of the challenges of manufacturing and stockpiling for a potential emergency situation (see Box 6-1).
Prioritization (who is treated first in an emergency when the supply of commodities is limited) is one of the toughest issues to resolve, Venkayya said. As an example of scenario-based planning and prioritization, Vankayya said that in 2007, the U.S. government published guidance for who would receive pandemic influenza vaccine first.3 This was a very hard, but necessary, set of discussions, Venkayya said. He continued that in an actual pandemic these decisions will be revisited, but they serve as a good starting point.
Priorities for Preparedness
Venkayya emphasized the need for a framework for global ongoing emerging infectious disease risk assessment, and collective agreement on medical countermeasures (MCMs) prioritization, target product profiles, and supply requirements. He reiterated support for designation of accountable parties and development of multipurpose solutions such as platform technologies. Venkayya also called for development of global, platform-specific concepts of operations for scaling up manufacturing using a network of manufacturers, contract manufacturing organizations, and publicly funded infrastructure. Predetermining the legal, regulatory, and policy
2 See http://www.phe.gov/Preparedness/mcm/phemce/Pages/strategy.aspx (accessed October 30, 2015).
3 See http://www.flu.gov/images/reports/pi_vaccine_allocation_guidance.pdf (accessed October 30, 2015).
Manufacturing and Stockpile Considerations
Product Development and Stockpile Priorities
- known unknowns versus unknown unknowns
- prioritization of threats once identified
- a “one bug–one drug” approach is not feasible; broad-spectrum or platform approaches are needed
- impractical to stockpile MCMs for large proportions of world population
- challenge of engaging industry given uncertainty of threat and market
Manufacturing and Logistics
- different products require different manufacturing platforms
- infrastructure, experience, and time to bring new capacity online for biologics (i.e., vaccines)
- uncertainty of demand
- shelf life of stockpiled products necessitates ongoing investment (e.g., maintaining/funding a warm manufacturing base to produce large quantities of pandemic vaccine)
- regulatory and legal frameworks
- ensuring equitable access to MCMs
- prioritization of populations to receive limited-availability MCMs
- policy maker/public appetite for investing in “insurance policies” for potential unknown health threats
SOURCE: Venkayya presentation, August 20, 2015.
framework that will support swift evaluation, licensure, and deployment of novel medical products is also essential, he said.
Ella suggested that small-scale contract manufacturing be set up for 10 candidates that are predicted to be likely to cause a pandemic. Planning, quality control standards, and stability studies can all be done ahead of time so that manufacturing can be started when needed. The manufacturing capabilities are there, he said, and the lines can be readily switched from one product to another. The prioritization of 10 potential candidates for which the global community can prepare is still needed.
Outbreaks have the greatest impact on the most vulnerable populations, which are the least likely to have access to critical health infrastructure and interventions. Addressing inequity is most difficult in a crisis, when policy makers and political leadership are incentivized to act strictly in their own national interest, and it is extraordinarily difficult for a political leader in any country to give away potentially lifesaving tools to other countries. The time to plan and put the necessary access framework in place is now, Venkayya stressed, right after an emergency and before the next one. He added that a global access framework should be a required part of a multilateral product development strategy.
Venkayya said that for influenza vaccine, the World Health Organization (WHO) has established a framework to ensure that industry commits a certain proportion of its capacity to developing countries. WHO asks influenza vaccine manufacturers to make cash donations on the basis of their annual sales of seasonal influenza vaccine, to donate vaccine, to commit to providing a certain amount of vaccine at an affordable price, or other options. This will not meet the full need in developing countries, but it is an interesting first step, he said, adding that the approach needs to be expanded to other threats. Venkayya noted that fulfillment of obligations must occur in real time (e.g., a 10 percent commitment means that 10 percent of every month’s production goes to the stockpile; a manufacturer should not hold off and give the last 10 percent of production, many months later, to the stockpile).
Hall said that governments nationalize supply from time to time, even when there is no outbreak. This is less of an issue when there is a diverse supplier base. She suggested that WHO should have the responsibility to decide where available vaccines should be deployed based on epidemiology, and it should be made public if countries nonetheless decide to nationalize their supply. Yamada also raised the issue of whether companies should be required to commit product to supply clinical trials.
Yamada pointed out that dedicated pandemic manufacturing capacity would be sitting idle most of the time. Ella agreed and said that large companies can sustain this, but it will be very difficult for small and medium-sized companies to sustain idle production capacity. Marks explained that, as part of a multipart deal between GlaxoSmithKline (GSK) and Novartis, half of the space in one of the GSK facilities is being dedicated to creating vaccines for emergent uses. This is an end-to-end unit with capabilities from vaccine design through manufacturing. The details are still in develop-
ment, but this dedicated facility is embedded in a research and development organization, and is not an isolated piece by itself, affording the ability to respond when needed. Venkayya said that the Biomedical Advanced Research and Development Authority (BARDA) has established partnerships with three companies that have vaccines and biologics manufacturing and fill finish capabilities so that, once a vaccine is developed, BARDA can quickly contract with those companies to scale up the manufacturing of that product. These sites are in use for other products that the companies are making, but they are available to BARDA should the need arise. Stoffels said that production technology solutions can enable greater output in smaller spaces, allowing for production capacity that can stay idle because it will be only, for example, a $50 million plant instead of a $500 million plant. Yamada pointed out the need for government funding to subsidize such capacity. Mahmoud raised the concern that every vaccine is different, and it is not as simple as making vaccine X one day and vaccine Y another. He suggested that the focus should instead be on expanding the global capacity to make vaccines. Levine pointed out that there are other limiting factors and highlighted the importance of developing new technologies. The reliance on egg-grown influenza vaccine, for example, is affected by the current outbreak of avian influenza, which has led to mass mortality in chickens.
Hamburg said that more sophisticated, more efficient, and higher-quality manufacturing processes are needed for drugs than the current batch-processing approach. Continuous-manufacturing approaches would benefit routine production and would be highly beneficial in situations where rapid scale-up is needed in response to emergencies. New technologies to advance manufacturing of therapeutics is greatly needed, especially with biologics becoming more prominent. Yamada lamented that research on manufacturing technologies is not funded by the National Institutes of Health, and industry does very little research in this area. He and Hamburg agreed that this is an area of national need.
David Ripin, Executive Vice President of Access and Malaria and Chief Scientific Officer of the Clinton Health Access Initiative (CHAI), discussed on-demand manufacturing as an alternative to stockpiling. He also noted the potential of the continuous-manufacturing approach suggested by Hamburg, but said that facilities designed for continuous production of a given product tend to be less suitable for other general use. He observed that, for small-molecule drugs, there is a massive capacity globally to make active ingredients and tablets in multifunctional facilities. He pointed out that the world is moving toward just-in-time supply, and the capacity to switch over a manufacturing line to respond to an emerging need relies on the availability of buffer stocks and other supplies. A just-in-time manufacturing approach could also create vulnerabilities if switching a line to
manufacture the new product leads to a stock shortage of other key products that might also be needed. Ripin said there is a financing cost to suppliers carrying a larger stock of the products they sell, and suggested that programs should be paying that cost as a mechanism for reserving some of the available multifunctional capacity.
Pauwels emphasized the complexity of building manufacturing lines for diagnostics. He added that most diagnostic companies are developing products for the developed world and do not take issues such as availability of cold chain into account.
Hall said that the United Nations Children’s Fund (UNICEF) has developed capacity in supply and logistics to meet the needs of its own programs in 190 countries, as well as to support requests for assistance from governments for their own national procurement programs. UNICEF stockpiles around $220 million worth of supplies in 202 locations across 63 countries, and around one-third to one-half of the stockpile is for emergency response to meet UNICEF’s commitments to support governments in UNICEF’s programmatic areas. These commitments guide what is put into inventory at the country level, while a planning process with the UN country team, nongovernmental organizations, and the government defines who will stockpile what, Hall explained. The UNICEF Global Supply Warehouse in Copenhagen is the world’s largest humanitarian warehouse. Hall noted that UNICEF does not stockpile any vaccines, instead working with companies to stockpile on UNICEF’s behalf. This is due to both regulatory issues and the irregularity of demand. UNICEF does inventory medicines, and is good distribution practice (GDP) certified and inspected by the Danish Medicines Authority.
Hall observed that the humanitarian system excels at preparation and rapid response, identifying target populations and investing in inventory for identified risks. She noted that the humanitarian and public health sectors are very segmented in their planning, and she suggested there could be more sharing from a stockpiling and deployment perspective. Whether it is UNICEF, a company, or a government, anyone carrying inventory must weigh the risks and decide on the “right” amount to store, she said. Hall concurred with Venkayya that there needs to be ownership of the process, and a plan behind it that accounts for the risk being taken. She suggested considering not only final, filled product for stockpiling, but also making agreements for stockpiling different stages of product completion. Ella agreed and suggested that storing bulk vaccine takes less space and extends shelf life. For most manufacturers, bulk manufacturing is the easier step and the limitation is fill capacity. Vaccine could be stored in bulk in country,
and filled as needed by a contract fill-and-finish organization. He noted that there are some issues of legal liability that will need to be addressed.
Ripin said that the ability to effectively and affordably maintain a stockpile relies on developing products that are stable and storable. He added that U.S. Department of Defense (DoD), working with U.S. Food and Drug Administration (FDA) and others, recognized that the shelf-life of some stable products in the Strategic National Stockpile (SNS) could be extended continuously, deferring the need to procure replacement product. FDA comprehensively tests and analyzes drugs and other medical material in the SNS. Products that pass testing are granted extended expiration dates, but they must undergo ongoing testing to monitor continued shelf life. Products that fail testing at any time are destroyed.
Supply chain encompasses everything it takes to get the product to the point of treatment, said Ripin. Recognizing the challenges faced in other areas will help inform product design, which Ripin said also includes studies of how the products can be most effectively deployed.
Lessons from Other Responses
A key lesson from the Haiti earthquake response, Ripin said, is to work with locally present groups. The philosophy of CHAI is to augment the capacity of the groups already present in a location. In Haiti, one of the strongest emergency relief partners in place was Partners In Health, which asked CHAI to manage supply for them, and that was the unmet need that CHAI worked to fill. Ripin suggested that community engagement is critical. It is not just a matter of educating the community, he said, but the community can also be a partner in delivering care. The delivery and deployment of products will include community care workers, lay providers, and untrained health workers. The Ebola response again demonstrated the importance of working with groups that are locally present.
Ripin also said to put local government in a leadership role, and to give the local leaders the tools and support they need to deliver care. There is a large infrastructure of health care delivery capacity in developing countries, but, Ripin said, we will have to accept that it might not look exactly the way that we want it to, and some of the care is not going to be delivered in the same way that we would necessarily choose to deliver it. Especially critical for international partners, Ripin said, is to be flexible and humble, and do whatever is needed, even if it is not necessarily the most attractive part of the response.
Ripin reiterated the point that there is a higher probability that a pan-
demic will emerge in locations that have more fragile health systems. As such, it is important to plan for products that can be delivered within that context. It is also important to remember that when a new infectious disease emerges and takes center stage, existing disease challenges still persist in the background and can have unrecognized costs. For example, malaria did not subside during the Ebola response. Ripin suggested that many cases of malaria likely went untreated during that time as people were fearful of going to clinics.
Supplying the Basics
Much of the discussion of stockpiling and rapid delivery has focused on vaccine and therapeutic products, but Ripin stressed the need to also ensure that basic supplies are in place. Regardless of the known or unknown threat, the global health community ought to have sufficient stocks of surgical gloves, basic masks, and basic protective garb, Ripin said. Even basic medical products such as intravenous fluids were a challenge to procure during the Ebola outbreak, he said. Earlier in the workshop, participants discussed the local stigmatization of patients and care providers during the Ebola crisis, but Ripin said that global stigmatization had a significant impact as well. As mentioned by Chan, closing borders and restricting travel was counterproductive and affected the supply chain of critical products, both interventions and basic supplies. Ripin noted that, despite the incredibly high mortality from Ebola at the beginning of the epidemic, only 2 of the last 14 cases of Ebola in Liberia died (one before getting to a health center and the other shortly after). The 12 that survived did so without any of the potential new investigational medical products (e.g., ZMAPP), and Ripin attributed their survival to a better care experience and access to basic medical supplies and diagnostics.
Delivering Supplies in a Crisis
UNICEF’s support to national Ebola programs helped to provide supplies for primary health care facilities, treatment and holding centers, and community care centers, as well as supply kits for households and schools. The composition of what UNICEF provided to each service delivery point varied by country. Hall pointed out the iterative nature of the supply needs, requiring supply chains to adapt and repurpose products as the program evolved.
In delivering supplies in Sierra Leone, Hall said, the definition of the need was the most challenging part of the supply chain because programs and facilities did not know what was needed. This included deciding which personal protective equipment (PPE) components were suitable for which
service delivery locations (e.g., Ebola treatment units versus community care centers, which Hall noted often became places of triage and early isolation due to bed shortages in the treatment units). Procurement was one of the easier parts of the supply chain, Hall said. Once they were able to forecast the PPE need, UNICEF increased its PPE suppliers from 2 to 17. Because UNICEF has established relationships, delivery and clearance was not an issue even when trade was restricted. Hall said that UNICEF led an air bridge, booking at least one charter flight per week into each of Guinea, Liberia, and Sierra Leone. The logistics of moving the supplies to all the service delivery locations was enormous, Hall said. UNICEF continually assessed consumption levels and aggregated the data to develop country-specific demand scenarios.4 She noted that Sierra Leone was well positioned for distribution efforts down to the district level as a result of a previous government health care and health supply chain initiative; however, this was not the case in Liberia and Guinea. Hall added that it is not just about delivering a supply. Health workers and burial teams need to be trained in use of the supplies, such as PPE. Awunyo-Akaba also stressed the need to consider the country-level issues for supply chain, such as the difficulty of terrain.
Yamada highlighted the many other players that have a critical role in a supply chain and that could have a significant impact on the rapidity and the cost of addressing a global emergency. These include, for example, international air freight service, trucking service, fuel suppliers, and others. Hall noted that for many of UNICEF’s humanitarian responses they have received donated air cargo. In the case of Ebola, they prebooked 10 charter flights over a 3-week period in order to negotiate the price down. Pauwels, Yamada, and others also discussed the critical role of the international military in deployment of facilities, services, and capabilities in a crisis. Armed forces in the impacted countries were also important in supporting logistical and operational functions.
Hall emphasized the need to consider demand scenarios when discussing stockpiles and supply. Yamada said that predicting demand is difficult because it depends on pathogen-related factors such as the nature of the illness, the rate of spread, and the case fatality rate. But perhaps there could be predictive models of how intervention with MCMs would be done for different pathogens. Ripin pointed out the need to distinguish need from demand (i.e., what the health system actually has the capacity to deliver, regardless of the number of cases). To some extent, the maximum capacity
of health systems to deliver care is predictable and might represent more of the true demand (as opposed to the need, which is dependent on the rate at which an epidemic is expanding).
Venkayya said that the SNS uses a scenario-based planning approach. The detailed analysis of numbers of cases, geographic distribution, severity of illness, and the comprehensive set of resources that would be necessary to take care of individuals in a chemical, biological, radiological, or nuclear event was very helpful because it exposed gaps in the upstream supply chain (e.g., for ventilator tubing, antibiotics, and normal saline). Venkayya stressed the value of bringing together a core group of partners to do comprehensive scenario-based planning for multicountry true pandemic events. He noted that the Ebola outbreak was geographically limited, and there was relatively free flow of commodities. However, a multicountry MERS outbreak might result in some sporadic border closures, and there would be a significant demand for PPE and the raw materials necessary to make PPE, and there are innumerable places where the supply chain could come up short and the response would suffer.
Bell added that not fully understanding demand can be a barrier when countries are stepping up to make donations of products and supplies. There should be some way of identifying the need so that donations can be deployed in an efficient manner.