Dispute over Pneumococcal Vaccine Initiative: A Response

An article by Ann Danaiya Usher in the December 5 edition of the Lancet focuses on aspects of the Advance Market Commitment pilot for pneumococcal vaccine that appear to be causing confusion. The article is similar to one published by the author in Development Today, a publication that has issued a series of negative (or at least skeptical) pieces about the AMC over the past few years. In particular, the article highlights the questions of whether the price offered for the first doses sold under the AMC subsidy program matters for the eventual health benefits anticipated from the AMC.The short answer is that it doesn’t.Understandably, that short answer leaves some people scratching their heads. In the Lancet piece, Usher seems to suggest that this confusion may be evidence of non-transparency on the part of GAVI. That’s not the case. Rather this is a complex business and it is necessary to spend some time diving into the details (all available on the AMC website), and understanding where judgement calls played a role. As a co-chair with Michael Kremer and Alice Albright of the CGD working group on AMCs that produced a report proposing a design for the AMC, and as a participant in some of the more recent steps, I feel compelled to offer fuller explanation than is included in Usher’s reports. I apologize for the length of this post, parts of which I suspect will interest only a small number of people who follow the AMC closely. (Disclaimer: These are my own views and do not represent those of the GAVI Alliance, which has the mandate to implement the AMC.)The BasicsThe Advance Market Commitment is an arrangement whereby donors make a binding commitment to buy a future vaccine, if and when it is developed, made available in sufficient quantities and demanded by countries for their immunization program. The donors offer a subsidy for each unit sold, up to a maximum number, at a price that provides a commercially viable return on investments in R&D and manufacturing capacity. As a quid pro quo for being assured of a relatively high price for initial sales to the developing world, the manufacturer(s) commits to making the vaccine available thereafter at a relatively low price – in concept, around the marginal cost of production. In short, the AMC was designed to provide an incentive for faster development and production of vaccines to benefit those in developing countries, where otherwise the market “pull” would be insufficient; and it was intended to accelerate uptake of the new vaccine by assuring country-level and donor decisionmakers that eventually the price would be affordable. A Little Background on the ProcessThe CGD Advance Market Commitment Working Group, co-chaired by Michael Kremer, Alice Albright and myself, was started in February 2003 and ended in April 2005 with the publication of the report Making Markets for Vaccines: Ideas to Action. Two of the 24 members of the AMC Working Group (not four, as reported in the Lancet article) opted not to sign onto the report recommendations: Donald Light, who had several objections to the concept and analysis (summarized here); and Jon Andrus of the Pan American Health Organization, who voiced concerns about the likely exclusion of middle-income countries such as those in Latin America and the Caribbean from participation in the AMC. (Donald Light has been a prolific writer about the AMC since that time, and was a source for Usher’s article in the Lancet and several of the earlier Development Today pieces.)Some of the working group recommendations, and particularly the idea of finding a suitable candidate vaccine for application of the AMC idea, were taken up by the G8 Finance Ministers, who asked the World Bank and GAVI to follow-up with a variety of expert group processes. This included convening health experts to select the vaccine (which turned out to be one against pneumococcal disease), and bringing together economic, legal and business experts to figure out how to apply the AMC concept to the particulars of that product. Along the way, six donors signed up to support the AMC (Italy, the UK, Norway, Canada, Russia and the Bill & Melinda Gates Foundation) and together committed $1.5 billion.I was involved in some but not all of the follow-up steps. I participated in the Economic Expert Group, which issued a final report in April 2008; and co-chaired with David Fleming the Implementation Working Group, which issued a report in July 2008. All those reports are here.Going from Ideas to Action: Not Always a Straight LineFor pneumococcal vaccine, the R&D was largely done: a close cousin of the desired product was available and in heavy demand in the U.S. and other lucrative markets. So the goal was to provide an incentive to one or more manufacturers to invest in the production capacity necessary to fulfill anticipated demand in low income countries, in a context where demand was expanding in high and middle income markets and competition was limited because a small number of manufacturers had or were likely to have a product to sell.Because of the particulars of pneumo, the Economic Expert Group recommended several modifications from the original concept proposed in Making Markets. These included:

• A supply commitment: That firms would be required to make a commitment ex ante to dedicate production capacity to supply a specific share of anticipated annual long-term demand in low income countries, and in return would receive a corresponding share of the AMC “pot” of $1.5 billion. (Long-term demand means the number of doses ordered once many countries have introduced the vaccine into their immunization programs and expanded its use. This can take several years.)
• A demand guarantee: That donors, via GAVI, would guarantee to purchase a portion of the anticipated demand even if it didn’t materialize from countries. GAVI as AMC implementer, would take on some of the demand risk.

Without those modifications, due to limited competition among suppliers, there was a possibility that firms would build production capacity just for the high and middle income countries, and would sell whatever modest surplus they happened to have under the AMC terms. This wouldn’t accelerate access much at all and might exacerbate a shortage situation for low-income countries. The demand guarantee was introduced to give some assurance to firms of GAVI’s confidence in its demand forecast and to reduce the demand risk for firms. The rationale for and descriptions of each of these modifications can be found in the Economic Expert Group final report.Now (finally!) comes the matter of price, and by this I mean the price that the supplier receives for each unit sold when the AMC subsidy is in place; this is referred to as the “AMC price.” This price is set ex ante and is not based on supplier-specific negotiations. It is a combination of the tail price (the amount that the supplier guarantees to sell the vaccine at after the AMC subsidy is used up) and the AMC subsidy. So if the AMC price is set at $7 per dose and the supplier offers a tail price of $3.50 then the subsidy is $3.50. If the supplier offers a tail price of $2 then the subsidy is $5 for a $7 AMC price. What is the “right” AMC price? According to Usher’s article, this is the subject of much confusion, with a critique that the price was not derived from a “transparent” formula or published quantitative analysis. Of particular interest is whether there is a difference in the magnitude of health benefits if the AMC price were $10 for each of the early doses (as she reports was on the table at one point in the donor discussions) or $7 per dose, which was recommended by the Implementation Working Group and accepted by AMC sponsors.The answer to that question is that the price itself doesn’t affect the number of doses obtained through the AMC, and so doesn’t affect the potential health benefits. Let’s work an example to show why. Say the long-term demand forecast is for 200 million doses per year. So a vaccine manufacturer comes along and says that it will build capacity to supply 50 percent of the estimated demand (scaling up to 100 million doses per year) in return for half of the AMC pot, or $750 million. After the AMC subsidy is used up through sales, the price drops to the “tail price” of, say, $3.50.At that tail price, if the AMC price is $7.00, the subsidy is $3.50, and the firm gets its $750 million after the first 214 million doses or so are purchased ($750m/$3.50=214m). As use of the vaccine is expanding from a very low base, it may take several years to get to that volume of sales. From that point on, it has to sell for $3.50 per dose. If, on the other hand, the AMC price were to be $10.00 per dose, then the subsidy is $6.50, and the firm gets its $750 million a bit sooner, after the first 115 million doses are sold ($750m/$6.50=115m). But in the end, the manufacturer gets the same amount regardless of AMC price, the same AMC pot buys the same number of doses, and that translates into a corresponding reduction in the incidence of pneumococcal disease. (There are some small nuances to take inflation into account, but I’m going to spare you those.)Obviously there might be a somewhat greater incentive for firms if they were to get the revenues sooner rather than later; they might rather have a higher AMC price than a lower one, all else equal. But under any realistic scenario the firms get their share of the AMC pot within a relatively short period, and so any difference in the strength of the incentive is small.This all begs the question of what the “right” AMC price is: what’s the price that’s going to get the “right” number of suppliers in the game, faster than they would otherwise turn their attention to the developing world? To sort this out, there is only one hard number to hold onto to get one’s analytical bearings: The marginal cost of production of the vaccine, including the capital investment. For obvious reasons, no manufacturer is going to sign up to a contract in which it is guaranteed to lose money on every dose sold.One can make some educated guesses about what that production cost is, recognizing that there are a very large number of factors involved; one manufacturer may have a lower cost process than another, and efficiencies can be obtained (or lost) over time. And those educated guesses certainly went into the decisions about the AMC price. But so did many other factors, including how various options were likely to be perceived by firms and by the general public in countries putting up some of the cash. In the end, as in all policy debates, those with skin in the game had to make the call in the face of imperfect information and the risk of criticism. From my perspective, that doesn’t translate into a lack of transparency, but rather is just an example of the fact that there are no black-and-white answers for some questions.The AMC development process has been a real education for me. One of the most important things I’ve learned is the value of personal commitment by the policymakers involved. There were innumerable moments when representatives of the donors, GAVI or the World Bank could have walked away, frustrated by the time required to meet, to understand, to negotiate and to agree. There have been multiple attempts by a relatively small number of critics to make participation in the AMC politically costly, and those continue. Economists from Finance Ministries have had to learn about the strains of pneumococcal disease, while public health specialists have had to watch demand and supply curves being drawn and redrawn until everyone in the room knew which one curved in which direction. None of it would have happened without the dedication of a relatively small group of people chipping away at the general AMC concept to shape a workable arrangement for a specific business and public health problem.For better or worse, there is more work ahead. The implementation of the AMC requires a major effort by GAVI, working closely with UNICEF around procurement issues. Those of us who have been involved closely or from the periphery are holding our breath, waiting to see whether supply agreements get signed and, eventually, life-saving vaccines roll off the filling line. And there are outstanding questions that affect not only the success of the pneumo AMC but the broader concept of accelerating introduction in low income countries of new and more expensive vaccines: What is the “affordable” price over the long term, particularly given the entry of multiple products (Hib, hepatitis B, pneumo, rotavirus and others)? Is the donor community willing to use its resources to both “prime the pump” through a mechanism like the AMC, and sustain support over the long term? As the pneumo AMC is implemented, and assessments are made about using an AMC-like arrangements for other vaccines (or other technologies) in the future, these are questions that need the same type of intensive analysis and debate as went into the AMC design itself.