The Cost of Being Sick: H1N1 and Paid Sick Days

Mr. Chairman and members of the Committee, I wish to thank you for the invitation to appear before you today to address issues related to our preparedness for H1N1 flu. While this influenza is, so far, proving less virulent than once feared, it is still a very dangerous virus.  This is especially true for vulnerable populations such as pregnant woman , young children, and those with compromised immune systems or lung disease.    H1N1 infections are expected to decline in November and December of 2009 but then peak again with higher mortality from March to May 2010. In this respect, some experts believe H1N1 may emulate the 1957 pandemic – decreasing late this year only to pick up again in the spring.

The good news is that we were much better prepared to deal with this flu than we would have been as recently as five years ago. This owes to steps taken by the current administration to contract for development of an H1N1 vaccine early last spring, when the virus first emerged. Collaborative steps to speed vaccine production were undertaken immediately, even before it was clear a vaccine would be needed. This includes work between U.S. government agencies, international partners, and drug firms to provide viral reference strains and reagents needed for vaccine production. These tasks were accomplished in record time despite technical challenges. In addition, extensive pandemic planning undertaken by the Bush Administration  left us with much better capacities to deal with this crisis.

But there are still significant gaps in our preparedness, and nagging vulnerabilities.

Too many of the policy choices we were confronted with in this crisis forced us to sacrifice on the speed and reliability of vaccine production in order to assuage concerns about vaccine safety. Vaccine supplies are increasing, but we still do not have the quantities we expected, in the time frame that we needed.  Among other things, we chose to forgo the use of vaccine additives that could  boost effectiveness and might have helped us stretch our limited supply of vaccine raw material over more shots. We are compelled to rely on old, unpredictable manufacturing technology because we haven’t developed the necessary capacities with more modern tools. We also lack domestic vaccine manufacturing facilities. In at least two cases we know of, this put the U.S. behind other countries in getting vaccine orders filled.

Moreover, while manufacturing problems at the drug firms contributed to delays in vaccine availability this year, the bottom line is that the policy choices we made also played a role. The drug makers are easy targets in our political culture and have recently received the brunt of official criticism from some public officials. But fault for today’s shortages don’t rest with them alone, any more than it rests with the public health officials overseeing our pandemic response. These are problems of biology and technology.

The bottom line is we have relied for too long on outdated capacity for our flu vaccines, in part because of our cultural reluctance to embrace new methods. This is not simply a regulatory issue, but also reflects the public mood when it comes to vaccine products.

There are good reasons why the regulation of vaccines is distinct. Vaccines are given to millions of otherwise healthy people, and administered over a compressed time period. This is especially true for flu vaccines. That rapid and widespread administration limits the ability to uncover “latent” risks after products are approved and marketed. It means that, by the time we intervene to prevent exposure to an emerging side effect, millions of people might have already received a seasonal product. This is a unique risk. For these reasons, a strong pre-market regulatory process is imperative. New vaccine technology, like any innovation, invariably brings some new uncertainties – heightening regulatory caution.

For all of these reasons and many others, we are slow to embrace change to flu vaccine production. But with the right tools and investments, we should be able to mitigate any reasonable risk. We can have more effective vaccines, and more predictable and timely supply, while maintaining our high degree of safety. This should be our focus.

These issues are matters of national security. The fact is that European countries share our regulatory standards and our focus on vaccine safety. But they are far ahead of us in using new and more reliable technology into their production of new flu vaccines. It’s true we remain farther ahead with other vaccine products, such as our adoption of conjugate vaccines or live attenuated approaches. But when it comes to pandemic planning, and response to flu, there is more we need to be doing.

Use of Vaccine Additives to Improve Yield and Effectiveness

One step to improving our readiness for the future is to better integrate the use of vaccine additives called adjuvants into our pandemic planning.

An adjuvant is a substance incorporated into a vaccine that enhances or directs the immune response of the vaccinated patient. Adjuvants are designed to bring the vaccine’s antigen into contact with the immune system. This works to enhance the magnitude of the immunity produced by the vaccine as well as the duration of the immune response.

Adjuvants have been used in vaccines for decades. The most commonly used adjuvant, aluminium salts, is used in some currently marketed vaccines, including those against diphtheria, tetanus and pertussis. The concept of embedding additives in vaccines that boost the immune response generated by the shot is actually quite old. Current manufacturing processes place a heavy emphasis on purification processes. But in the old days, vaccines were far more “reactogenic” - meaning they caused a lot more local reactions. Those reactions (typically sore arms) were the results of the impurities in the vaccines. These impurities were effectively acting as adjuvants by eliciting local immune reactions that boosted the effects of the vaccine by recruiting immune cells to the site of the vaccination. This worked to more effectively present a vaccine’s antigen to the immune system.

Novartis  and GSK, among other drug firms, have done innovative work incorporating new generations of adjuvants into vaccines marketed in Europe this fall for H1N1. A lot of the recent activity in Europe to deploy adjuvants was based on “mock up” preparations of pandemic vaccines that those nations had been pre-approved and stockpiled.

When H1N1 first surfaced last spring, the World Health Organization asked countries to use “antigen sparing” approaches to pandemic vaccination. Using an adjuvant is one of the main antigen sparing options available. Adjuvants can work to increase the protective effects of a given quantity of vaccine. In the U.S., our decision to forgo their use limited our ability to stretch our already limited stock of H1N1 vaccine raw material (the antigen).  It is worth noting that no country has had earlier large supplies of vaccine, including in Europe. The three countries first out with substantial vaccine (the U.S., Australia and China) all used non-adjuvanted egg based vaccines. So the capacity challenges are a global problem. Merely adopting adjuvants does not ameliorate them. There is no single solution. But to improve for the future, we need to embrace methods that can improve vaccine supply and effectiveness.

In 2008, GSK became the first company to obtain a European license for an adjuvanted prepandemic vaccine, Prepandrix. This vaccine is designed to raise immune protection against several strains of the H5N1 (Avian) flu virus.  GSK also recently became the first drug manufacturer to get U.S. Food and Drug Administration (FDA) approval for a modern adjuvant that is used in conjunction with a vaccine distributed domestically. That vaccine, Cervarix is administered to prevent cervical cancer and precancerous lesions caused by human papillomavirus (HPV) types 16 and 18. Cervarix contains the adjuvant ASO4, which is a combination of aluminum hydroxide  and monophosphoryl lipid A (MPL).  It is the first vaccine licensed by the FDA that includes MPL as an adjuvant. ASO4 is a close cousin of the adjuvants that are already in wide use in Europe, and shares some similarities  to adjuvants included in some of the versions of H1N1 vaccine being used around the world.

There is no adjuvant approved for use in a flu preparation in the U.S. and no adjuvanted H1N1 vaccine available in this country.

Integrating an adjuvant into the U.S. H1N1 vaccine would not have been as easy as borrowing the data used by Europe. For one thing, the European approvals for pandemic vaccines, and most of the clinical data that were reviewed by the European Medicines Agency (EMEA) to support them, did not study the identical vaccine antigen and are not manufactured at the same facilities from which the U.S. H1N1 vaccines. There are differences that potentially can occur when different antigens are mixed with different adjuvants. So it’s not a sure bet that the antigen available for the U.S. vaccine could be effectively used in conjunction with the same adjuvants being used in the European vaccines.

The safety profile of vaccines can also be affected by minor changes in how a protein is presented to the immune system, as occurs when an adjuvant is included in a vaccine. So one can’t reflexively conclude that the antigen available for the U.S. vaccines would have been easily adapted with an adjuvant just because the European antigen was amenable to this process. Nonetheless, there is good reason to believe that for most patients, these adjuvants (one is already used in a U.S. stockpiled vaccine that targets pandemic avian flu ) could have been adapted to the U.S. shots, and would have boosted our present supply of a H1N1 vaccine as much as fourfold.  Even more shots could have been produced for kids, since data shows the adjuvanted vaccines require even less antigen when used in children.      

U.S. public health authorities didn’t dismiss adjuvants outright. Quite the opposite. FDA, CDC, and National Institutes of Health (NIH) laid some groundwork last spring to evaluate the adjuvants in the event that the H1N1 vaccine proved to be ineffective in the absence of these components. It was with the strong urging of the FDA that studies by vaccine manufacturers and NIH included both adjuvanted and non-adjuvanted formulations of H1N1 vaccine. The Department of Health and Human Services (HHS) also purchased and filled and finished a large stockpile of adjuvant in case it was needed.

In addition, U.S. public health authorities asked for data that could inform the effects of adjuvants and whether they would be beneficial and needed for H1N1 vaccine. In fact, the studies that regulators around the world relied on to evaluate the immunogenicity of both non-adjuvanted and adjuvanted vaccines are largely the result of requests for this data by FDA. The U.S. worked to keep an adjuvant option “on the table” was it to be needed.

Yet despite the foundational work done by FDA and others, the U.S. might not have been prepared to license an adjuvanted H1N1 vaccine through our customary regulatory process should it have been necessary. In all likelihood, if we had to incorporate adjuvant this fall, we would still have been forced to make an adjuvanted H1N1 vaccine available under an Emergency Use Authorization (EUA).  This is an authority that authorizes use of a product for treatment or prevention of well-defined, public health emergencies when the FDA hasn’t already approved the relevant product for this specific use.  A vaccine supplied through such an expedited authorization such as a EUA would have surely raised public concerns about its safety testing, perhaps reducing vaccination rates and offsetting any public health gains achieved by the use of the adjuvant. As a result, while the option of using an adjuvant was kept on the table by U.S. authorities, it was set on the very edge of the table.

Ultimately, the U.S. decision to not employ adjuvants was based on clinical data that showed an excellent response to standard doses of the licensed vaccines in the absence of any adjuvants. But that meant that the H1N1 vaccine required much higher quantities of vaccine raw material (antigen) than would have been required if adjuvants had been incorporated.    While the amount of antigen in the U.S. H1N1 vaccine is equivalent to the quantity used in the seasonal flu vaccine distributed around the world each year, in this case, we had very limited quantities of H1N1 antigen. Stretching supply was imperative. In the U.S., we were compelled to spread a limited supply of vaccine antigen across fewer shots than Europeans.

In a future pandemic, we may not have this same opportunity. Even today, the decision to forgo the use of adjuvant has to be considered as one of the tradeoffs contributing to our current H1N1 vaccine shortage. This kind of tradeoff doesn’t need to exist in the future.

What measures can be taken to improve our process for evaluating vaccine adjuvants?

First, FDA should consider creating formal guidance on the development and use of adjuvants to help guide product developers. The EMEA developed formal guidance on adjuvants three years ago. The document is available on that agency’s website.  FDA doesn’t have a similar guidance document, and while it hasn’t indicated it plans to write one, the FDA held a meeting on the topic in December 2008. It workshop could serve as a prelude to the development of formal guidance-writing process.

The U.S. should also consider stockpiling pre-approved vaccine preparations that could be used in a public health emergency. There is now ample experience in Europe on which we can draw.  Adjuvants are not approved as stand alone substances because they do not always perform the same with different vaccines or types of vaccines or, at times, even with different versions of the same antigen.  Nonetheless, the European strategy of having pandemic vaccines pre-approved, as mock-ups, was a prudent step.

Upgrading our Manufacturing Technology

Seasonal flu vaccines and the H1N1 vaccine are still made by the same process that has been used for fifty years: they are grown inside chicken eggs.  This process is unpredictable, slow, and difficult to scale. It is also expensive, costing more than $300 million to build a new plant and requiring more than five years to bring an egg-based production facility online.

Here is how the egg-based process works: Flu, as with any virus, will grow only in living cells. In the case of flu vaccine, production of the vaccine components has used the cells of embryonated (fertilized) hens’ eggs. The success of this system is primarily dependent upon the availability of adequate flocks of chickens. These flocks must be hatched about six months in advance to achieve maturity at the time that the eggs are needed. A bipartisan investment that helped improve our readiness was support of year round flocks. Nonetheless this egg-based process requires long lead times and has other risks.

The flocks, for example, are susceptible to their own diseases.  Another challenge of the egg-based process is virus yield. This refers to the number of viral particles that come out of an egg that could be used to make the vaccine. As a rule of thumb, one to three eggs are needed to produce each individual shot of the seasonal flu vaccine. Eggs are typically low-yield factories for the production of vaccine components.

This was certainly true this year. The H1N1 virus that was adapted by the Centers for Disease Control (CDC) for growing inside the chicken eggs, and sent to the manufacturers as the “seed” stock  (for jumpstarting manufacturing lines) was slow in being shipped to the drug firms owing to the difficulty in developing this template strain. Once it arrived, it was not well suited to the production lines, and yielded low quantities of vaccine antigen.    Manufacturers spent several weeks before they realized this seed stock was yielding low vaccine quantities. It took still more weeks for the drug firms to re-engineer the seed stock to come up with a more effective template for growing vaccine antigen in the chicken eggs.    This experience underscores the unpredictable qualities of our present flu vaccine manufacturing process, and how vulnerable we are as a result of our dependence on it.

Because of the uncertainties and delays inherent to this production process -- and because the emergence of pandemic strains of influenza virus may occur outside the normal time frame for vaccine production (when chicken flocks are not at peak availability) we need alternative production systems for flu vaccine. The principal alternative to the egg-based process is tissue culture cell lines that can be used as incubators for viral replication. 

Using cell cultures instead of chicken eggs cuts three to four weeks from the time required to mass-produce a vaccine. But the biggest advantage of cell-based manufacturing is its more rapid scale-up and is potentially better predictability. These attributes are typically more variable using older egg-based processes. Moreover, the use of hundreds of thousands of eggs can be a more dirty process, making it prone to production glitches.

There are many approved cell culture vaccines made in the U.S. This includes most of our viral vaccines such as Measles, Mumps and Rubella (MMR) as well as vaccines for polio and Zoster, among others. An issue for flu vaccines has been getting good yield and a good clinical response using cell cultures. Only in recent years has there been real progress on these steps. As a result, the U.S. has recently begun to scale up work on cell-based manufacturing for influenza vaccines. More needs to be done. Our current vulnerabilities are too significant to be satisfied with merely incremental progress.

The Biomedical Advanced Research and Development Authority (BARDA) awarded one federal contract for $487 million last spring to Novartis for the construction of the first U.S. facility to manufacture cell-based flu vaccine.  That facility is scheduled to open this year, but it won’t be producing licensed vaccine until 2014.    GSK and Sanofi-Aventis are also working on cell-based production of influenza vaccine.  Baxter recently became the first company to gain marketing authorization by the European Commission for a cell-based vaccine.  That cell-based vaccine product is not available in the U.S.

Cell based vaccine production is not without its own obstacles, and risks. In addition to issues around getting adequate yields from cell-based production processes, there are also challenges with immunogenicity  and reactogenicity . All of these problems have come up in past attempts to scale cell based production processes. There is also a remote and theoretical safety concern around the ability of genetic material to jump from the cell lines, into the vaccine, and then integrate into human tissues. FDA has issued a guidance to provide a pathway for safe use of novel cell substrates that tries to address the proper testing that flu vaccine manufacturers should undertake in order to rule out these risks.

Given the strategic advantages of the cell-based process, we need to invest in developing this capacity more quickly. BARDA should support development of similar facilities to the one being constructed in North Carolina. A typical cell-based facility costs as much as $600 million and would only be able to produce about 40 million doses of seasonal “trivalent” flu vaccine a year. The Novartis facility will be able to produce around 150 million doses of “monovalent” vaccine--containing just one viral strain, as opposed to the seasonal flu vaccine, which contains three different viral strains--in the event of a pandemic.

All of this illustrates the more challenging economics of vaccine production, for which significant upfront expenditures are required to build facilities capable of producing largely fixed capacities of vaccine. So long as seasonal flu vaccines remain commoditized products, with slim margins and little product differentiation (public health agencies want vaccines coming from different manufacturers to be largely interchangeable) then there will not be large enough private profits to support substantial new investments in manufacturing infrastructure. Getting additional facilities on-line will require federal investment. This capacity, however, is a matter of national strategic security and should be a U.S. priority.  

Ensuring Domestic Production Capabilities

We also need to make sure that an adequate proportion of the worldwide influenza vaccine production capacity is domiciled in the U.S. – enough to adequately supply a reasonable portion of the U.S. market in the event of a pandemic.

It is hard to envision other nations allowing limited supply of vaccine raw material to be shipped outside their borders in the event of a full-blown pandemic with a very dangerous flu. More likely, nations would take steps to nationalize their domestic production capacity.

The drawback to relying on foreign plants was made clear recently when foreign countries claimed priority for the H1N1 vaccine produced in their own countries. That was the case in Australia, where the government pressured vaccine manufacturer CSL to keep its vaccine at home instead of fulfilling its contract for 36 million doses of swine flu vaccine for the United States.      In Canada, where GSK maintains one of its two flu vaccine production facilities, the company had to assure the Canadian government that the Canadian population would be served first from that facility before any other countries that rely on that manufacturing site – including the U.S. – received fulfillment of their H1N1 vaccine orders.

This risk is compounded by the fact that all but one of the vaccine production facilities we depend on is located outside the United States.  There are five companies licensed to sell seasonal flu vaccine in the U.S. But only one, Sanofi-Pasteur, has a domestically located plant. The others — GlaxoSmithKline, Novartis, CSL Ltd. and MedImmune — presently use plants in Canada, Germany and Australia, to supply the U.S. market.

Much of the new technology is also being located outside the U.S. After the U.S. firm MedImmune was acquired by AstraZeneca, additional production capacity was located in Cambridge, UK in 2008. Novartis, based in Switzerland, operates a cell-culture vaccine production facility in Marburg, Germany. The cell culture facility maintained by Baxter for production of flu vaccine is located in the Czech Republic.

There also appears to be significant limitations in global fill and finishing capacities for flu vaccine. This also limits supply. In addition, concerns about trace amounts of the mercury-containing vaccine preservative thimerosol, found in multi-dose vials of flu vaccine, prompted public health officials to request drug firms manufacture more single-dose syringes. This took longer and added delays to vaccine availability.

There are lingering concerns that thimerosol is linked to autism, despite well conducted studies that show that the vaccine preservative is safe. If we are going to let these kinds of theoretical fears drive decisions about how vaccines are packaged, than we ought to invest in better finishing capacity and preservative-free technologies. Ideally, we also need more of the companies that produce flu vaccines to locate new filling and finishing facilities in the U.S.

One of the additional business impediments companies face to making investments in multiple, differently situated vaccine production facilities stems from how these facilities are regulated. The vaccine produced from each facility needs to be separately licensed by both the FDA and the EMEA. That means that if the same company produces flu vaccine at two different facilities (even in cases where it uses the same processes at each facility) the company often has to conduct separate clinical trials for each vaccine. While FDA has approved vaccines where little or no U.S.-specific data was available, there remain many situations where redundant trials were required or European data was not fully leveraged.

This drives developers to expand existing facilities rather than create new ones. Since the clinical trials require substantial investments of time and money, it is far more economical to maintain a few very large vaccine production facilities. After all, each facility’s vaccine will be treated as a completely new product with its own expensive clinical trials. There are good scientific reasons why biologicals coming from distinct facilities are treated independently by drug regulators. But there may be better ways to enable more cooperation between requirements set forth by different regulators or make use of studies that could bridge between products from a single manufacturer’s different manufacturing lines.

The ability to conduct these kinds of bridging studies, if they could streamline the requirements for entirely separate clinical trials, could save time and money. It would also reduce the economic impediments firms face to creating redundant manufacturing capacity.

Other measures that would help create more domestic capacity include guaranteed markets for seasonal flu vaccines. This would create additional incentives for building U.S. manufacturing capacity, especially if the tender process favored domestic manufacturers.

Other Areas for Improvement

We need to develop new types of vaccines. BARDA has made grants available to fund research into completely new platforms for vaccinating against flu. Just this past June, BARDA awarded a research and development contract for work on a recombinant flu vaccine. We are making incremental but meaningful progress. We should be undertaking a more robust process to put substantial resources behind these scientific efforts.

The complexity of developing a vaccine against pandemic flu is similar to the problems posed by development of the seasonal flu shots. The vaccine needs to be adapted to match each specific strain of the flu virus. In the case of the seasonal flu, we have to develop a new vaccine each year to guard against that season’s circulating strains of influenza.

It also means that we depend on just-in-time delivery when it comes to flu vaccine. This owes to the fact that the vaccine targets proteins on the surface of the flu virus that itself undergo easy mutation. Since these proteins change easily, a new vaccine must be developed to target the unique proteins found on each particular strain of influenza.

Better technologies can enable development of vaccines that require much shorter development timelines, or that protect against a broader range of flu strains.

On the first point, for example, Virus Like Particles (VLPs) have been suggested as a promising platform for new viral vaccines. In the light of a pandemic threat, VLPs have been recently developed as a new generation of non-egg based cell culture-derived vaccine candidates against influenza infection. 

Influenza VLPs are formed by a self-assembly process incorporating structural proteins of the flu virus.  These particles resemble the virus from which they were derived but lack viral nucleic acid, meaning that they are not infectious. VLPs used as vaccines are often very effective at eliciting both T cell and B cell immune responses. The human papillomavirus and Hepatitis B vaccines are the first VLP-based vaccines approved by the FDA.

Research suggests that VLP vaccines could provide stronger and longer-lasting protection against flu viruses than conventional vaccines.  Production may begin as soon as the genetic sequence of the virus is published online, without an actual sample of the agent, and it may take as little as 12 weeks, compared to 9 months for traditional vaccines.  The VLP may be grown in either plants or insect cells. As it contains no genetic material, some ingredients of traditional vaccines such as formalin and detergent treatments, are not needed.  In some recent clinical trials, VLP vaccines appeared to provide complete protection against both the H5N1 avian influenza virus and the 1918 Spanish influenza virus.

There is also opportunity to create a vaccine that protects against a broader variety of influenza strains, reducing the need to tailor a new vaccine to each individual strain of circulating flu. A universal vaccine would target more “conserved” regions of the flu virus’s structural proteins--parts of the flu virus architecture that do not undergo much mutation and, therefore, are unlikely to change, regardless of the particular strain of flu.

Right now, our vaccines target proteins that are on the outer surface of the flu virus. Since our immune systems attack these proteins, the proteins themselves undergo adaptation, mutation, and change in order to evade our immune response. But structural proteins that are core components of the architecture of all flu viruses would be less likely to undergo mutation, regardless of the pressure from nature to change in order to survive.

Theoretically, to target these core proteins, a universal vaccine would need to recruit our T cells to attack the flu virus, as opposed to today’s vaccines, which recruit an antibody response. For that reason, some suggest that such a “universal” vaccine would more likely be a therapeutic tool, as opposed to a protective vaccine. There is some literature to suggest that a T cell response alone may not be sufficient to protect us fully from flu, but work continues, and a universal vaccine is at least possible.

Drug firms sometimes complain that there is a disconnect between the advice and goals of different government agencies, especially between those charged with trying to develop new technologies (BARDA) and those charged with ensuring their safety (FDA).

It remains important for FDA to preserve its distinct mission to assure product safety and effectiveness and for the agency to remain independent. But when it comes to areas of critical public health need, where the government is engaged in a substantial effort to fund development of new technology, there’s more we can do. FDA meets early with academic and industry developers of novel technologies especially for critical public health needs like flu and terrorism. But there may be more opportunities to create clearer pathways to market by also engaging FDA more closely in the government procurement process.

One opportunity is to couple BARDA funding of new technology with regulatory programs that provide additional, early feedback to sponsors development those new methods. Multiple studies have shown that early and frequent FDA feedback helps sponsors avoid mistakes and results in timelier access to safe and effective products. This kind of regulatory effort is time and labor intensive, however, and would need to be funded inside FDA.

Finally, we also need to spend time examining how limited vaccine has been distributed during this pandemic, and take steps to put in place a better process for the future. My own view is that we should have relied more on the clinical community as a way to target the vaccine to high risk Americans. Doctors who treat high-risk patient populations – for example obstetricians that see pregnant women or pulmonologists who treat people with lung disease – in many cases had no access to the vaccine in many states. To target these populations of patients, we need to work through, and target, the doctors that care for them.

Conclusion

The Obama team deserves credit for ordering vaccines early last spring when H1N1 first emerged and for acting quickly to support their development. It wasn’t clear, at that moment, whether H1N1 would emerge as a pandemic or fade into the summer and fail to re-emerge in the fall. The Administration’s decision to undertake a crash effort to field vaccine saved lives.  Many of the shortcomings in our current preparedness are not the product of policy choices, but are challenges that relate to biology and the inherent complexity of targeting viruses that change rapidly and frequently. The fact that the U.S. has quickly fielded a program with high quality licensed vaccines despite the old technology and processes we relied on is a substantial public health accomplishment.

But these successes, and our prudence in fielding an early supply of H1N1 vaccine, shouldn’t obscure the fact that we also made deliberate decisions to rely on old methods rather than adapt new ones because of our concerns about safety and our comfort with the tried and true approaches. Some of these policy choices had consequences, contributing to our current shortages. These tradeoffs can be reduced in the future if we make a concerted effort today to increase our capacity for timely development of safe, effective and innovative flu vaccines.