Strategic Demand Forecast History

Strategic Demand Forecast History 2003 – 2008 May 2009 Table of Contents 1. Executive Summary ............................................................................................................2 Forecast Methodology .......................................................................................................... 2 Pneumococcal Vaccine Need and Potential Vaccine Market ................................................ 3 Pneumococcal Vaccine Strategic Demand Forecast (GAVI-Eligible Countries) .................... 3 Strategic Demand Forecast Refinements ............................................................................. 4 Strategic Demand Forecast Performance............................................................................. 4 Strategic Demand Forecast Ongoing Monitoring and Evaluation .......................................... 5 2. Introduction .........................................................................................................................5 3. Demand Forecasting Roundtable (2003) ...........................................................................7 4. Strategic Demand Forecast Methodology Overview ........................................................9 5. Strategic Demand Forecast v1.0 (2005-2006) ..................................................................11 A. Vaccine Need and Potential Market Size .......................................................................11 B. Potential Demand ..........................................................................................................15 i. ii. iii. iv. 1st Vaccine Availability ................................................................................................15 Earliest Time to Adoption ............................................................................................16 Time to Peak Coverage ..............................................................................................20 Results .......................................................................................................................20 C. Strategic Demand Forecast Impact ...............................................................................21 6. Strategic Demand Forecast v2.0 (2006-2008) ..................................................................23 7. Strategic Demand Forecast Performance .......................................................................28 8. Strategic Demand Forecast Ongoing Monitoring and Evaluation .................................31 Appendix A: 2003 & 2005 Demand Forecasting Roundtable Participants ..........................32 Appendix B: Key Forecast Variables with Sources and Vetting Groups ............................33 This document was developed by Lois Privor-Dumm and Orin Levine of the PneumoADIP and Sandy Wrobel, Daniel Machemer, Kimberly Swanson and Mary Story of Applied Strategies. We would like to acknowledge the efforts of Angeline Nanni and Andrew Jones, both in development of the strategic demand forecast and their contributions to this document. 1 1. Executive Summary In 2003, the GAVI Alliance (GAVI) funded the Pneumococcal Vaccines Accelerated Development and Introduction Plan (PneumoADIP) at Johns Hopkins Bloomberg School of Public Health to improve child health in developing countries by accelerating the development and introduction of new vaccines against pneumococcal disease. One of PneumoADIP’s goals was to ensure an affordable and sustainable supply of pneumococcal vaccines for GAVI-eligible countries . To enable vaccine suppliers, global donors, and countries to make informed decisions about pneumococcal vaccine supply, financing, and adoption, respectively, PneumoADIP developed a strategic demand forecast. Created for products that are typically 5 or more years from launch, strategic demand forecasts represent demand over the long-term, usually 5-20 or more years into the future. This timeframe is critical for evaluating the risk and return associated with large vaccine-related investments. PneumoADIP’s strategic demand forecast was developed with industry best practices in mind and a robust forecast development and vetting process to ensure a result useful to all key stakeholders, including vaccine suppliers, global donors, and country decision-makers. Prior to the PneumoADIP strategic demand forecast, supply chain forecasts spanning one to three years were primarily used by the global health community to forecast demand for global health products. While necessary for vaccine procurement decisions, these short-term forecasts were not designed or well-suited to help facilitate accelerated introduction in developing countries or motivate supplier investment in manufacturing capacity. 1 Methodology PneumoADIP, with the support of Applied Strategies Consulting, developed its strategic demand forecast methodology from the bottom up. The process involved multiple stakeholders who provided perspective of more than 25 donors, suppliers and countries. Two meetings were held to review the process and needs for the strategic demand forecast and multiple consultations including in-person visits with each of the WHO regional offices in 2003-2004 were held to develop assumptions that underlie the estimates of demand, Recognizing that there was little history in the public sector of strategic demand forecasting and inconsistent use of key terminology, an important initial step was to build a common set of terms and definitions to support the methodology and the communication of the forecast. The terms used to describe the strategic demand forecast are as follows: Vaccine Need, the amount of vaccine required if the total relevant population within a country was vaccinated with the required number of doses, provided the foundation for the forecast methodology. Recognizing that the entire relevant population is unlikely to have access to immunization services, vaccine need is adjusted by the representative vaccine coverage rate – DTP3 as the proxy for pneumococcal vaccines – to establish the Potential Vaccine Market. The strategic demand forecast takes potential vaccine market one step further to determine the Potential Demand. Potential demand takes into account the earliest time to adoption for each country as well as the country’s time to peak coverage once the vaccine is adopted. In general, potential demand assumes 1 GAVI-eligible countries are the 72 countries eligible to receive GAVI funding based on economic status (GNI per capita < $1000). 2 no financing or supply constraints and is, therefore, the demand we wish to achieve. Potential demand is the ultimate ambition of the strategic efforts to accelerate vaccine introduction. Potential demand also provides the foundation for Forecasted Demand. Forecasted demand accounts for available supply, global donor support programs, countries’ product preferences, and country vaccination program sustainability based on projected vaccine prices and a country’s willingness to pay. Willingness to pay is assessed as a country’s maximum acceptable vaccine price (MAVP). Assuming the MAVP is equal to or greater than the projected vaccine price, a country will be deemed able to sustain the vaccination program once global donor financing is no longer available. If a country is deemed unable to sustain its vaccination program, the country will not be included in forecasted demand. To support other program objectives, PneumoADIP went beyond the strategic demand forecast to determine forecasted demand given alternative supplier and financing scenarios. These additional analyses were very useful for assessing the implications of forecasted demand on supplier investment returns and global donor financing policies. However, the focus of this historical review is the potential demand (i.e., strategic demand forecast). Vaccine Need and Potential Vaccine Market For pneumococcal vaccines, the vaccine need in 2010 across all high-, middle-, and low-income 2 countries was forecasted at approximately 420 million doses, representing full vaccination of 136 million births worldwide. The potential global vaccine market in 2010 was forecasted to be just over 350 million doses. GAVI-eligible countries were expected to account for nearly 180 million of these doses, or just over 50% of the potential vaccine market. Applying expected public and private market vaccine price ranges to each of these market segments resulted in a potential annual global vaccine market value of approximately $5 billion, with high-, middle-, and low-income markets accounting for $2.4 billion (46%), $1.7 billion (33%), and $1.1 billion (21%), respectively. A market value in excess of $1 billion annually should provide suppliers adequate incentive to accelerate introduction of these vaccines in developing countries. Vaccine Strategic Demand Forecast (GAVI-Eligible Countries) The development of assumptions was a critical step in the process, and involved multiple stakeholders. The key assumptions included: • • • • • • • • Pneumococcal vaccines remain a GAVI priority Availability of GAVI & donor funding throughout the forecast period GAVI application process remains in place Co-pay is not increased above willingness to pay Countries receive adequate technical support and apply for GAVI funding There is adequate supply There are no serious issues with the vaccines, e.g., safety GAVI policies continue to allow all 72 currently eligible countries (i.e., including India) equal access to GAVI financing for the vaccine 2 High-, middle-, and low-income countries refer to the World Bank country income classifications based on GNI per capita: highincome World Bank classification (~$11,456 and above); middle-income refers to the upper-middle World Bank classification (~$936 - $11,455); low-income includes the low and lower-middle World Bank classifications (~$935 or less). 3 Based on the earliest time to adoption assessment, GAVI-eligible countries were categorized as either early, mid, or late adopters. PneumoADIP’s assessment resulted in 26 early adopter countries expected 3 to adopt within six years of vaccine availability to GAVI-eligible countries. Twenty-five countries were expected to be mid adopters (2016-2020) and 21 countries late adopters (2021-2025). Based on this adoption forecast and assessments for each country’s time to peak vaccine coverage, PneumoADIP calculated potential demand. By 2015, potential demand was forecasted at approximately 50 million doses, increasing to nearly 160 million doses by 2020, and peaking at just over 200 million 4 doses by 2030. Assuming pneumococcal vaccines prevent 7 deaths per 1000 children vaccinated (the impact observed in the Gambia trial), the 2005 demand forecast (v1.0) projects that approximately 5.3 million deaths can be averted between 2010 and 2030. Demand Forecast Refinements Following the launch of its Strategic Demand Forecast v1.0, PneumoADIP continued its efforts to refine the forecast based on evolving events and updated information. By late 2007, PneumoADIP decided to formally issue a strategic demand forecast v2.0 to reflect the many changes that had occurred in the two years since the launch of the initial forecast. The major events motivating this updated forecast included: • Accelerated country adoption and increased vaccine interest expected due to: – Increased understanding of disease burden – WHO/SAGE recommendation for pneumococcal vaccine use – Pneumococcal vaccine Pilot AMC and GAVI Phase 2 financing policy Updated underlying population and vaccine coverage rate forecasts • Accounting for these changes in expected country adoption, strategic demand forecast v2.0 estimated an additional 250 million doses of pneumococcal vaccine over the v1.0 forecast would be required in the 2008 – 2015 timeframe and an additional 475 million from 2008 through 2020. If supply can be made available, approximately 2.5 million additional deaths could be averted between 2008 and 2030 through the accelerated introduction of pneumococcal vaccine Demand Forecast Performance Strategic demand forecasts are intended to predict the shape of a market 5 to 20 years in the future. Because of their nature, these forecasts are uncertain and likely based on benchmarks, broad assumptions, and limited data. As a result, evaluations of their accuracy should only be conducted for the purpose of improving upon the methodology or assessment assumptions. Continual evaluation is critical to ensuring future strategic demand forecasts build upon the lessons learned by earlier forecasters. To develop these lessons learned, the PneumoADIP strategic demand forecasts were evaluated against currently available information. Since pneumococcal vaccines have not yet been introduced into GAVIeligible countries, there is relatively little data available. At this point in time, only vaccine availability and 3 Vaccine availability for GAVI-eligible countries was defined as the date of WHO prequalification, a prerequisite for the release of GAVI financing funds. The first pneumococcal vaccine for use by all GAVI-eligible countries was expected to be WHO prequalified in 2010. 4 Cutts, FT, et al. “Efficacy of nine-valent pneumococcal conjugate vaccine against pneumonia and invasive pneumococcal disease in The Gambia: randomized, double-blind, placebo-controlled trial.” The Lancet, 2005: 1139–1146. 4 country adoption assumptions can be evaluated based on actual submissions for licensure and approved GAVI applications. Demand Forecast Ongoing Monitoring and Evaluation Eleven GAVI countries have currently been approved to receive GAVI support for pneumococcal conjugate vaccine. Although adoption is underway, the longer-term strategic demand forecast will still play a critical role in helping key stakeholders make informed decisions in the face of an ever-changing market environment. Updated forecasts will be needed to encourage the development of new pneumococcal vaccines, including conjugate vaccines from emerging suppliers and the next generation protein vaccines. Ongoing monitoring and evaluation of the strategic demand forecast is also important for managing stakeholder expectations. If the underlying assumptions supporting the forecast do not pan out as expected, the implications could be significant. Semi-annual forecast evaluations and updates with implication analyses are highly recommended to support the most efficient adoption and uptake of this lifesaving vaccine. On a larger scale, as multiple vaccines become available simultaneously, countries will face tough decisions about which vaccines to introduce and when. Their decisions will need to take into account the trade-offs between disease burden, cost effectiveness, infrastructure requirements, potential product health impact, as well as other political and cultural factors. Strategic demand forecasts will need to account for the potential impact multiple new vaccine options will have on the decision to introduce a specific vaccine. 2. Introduction Historically, developing countries have waited 15 to 20 years to adopt vaccines after they have been introduced in industrialized countries. There are many explanations for this time lag – prices that are perceived as unaffordable, a lack of manufacturing capacity, perceived risk of limited demand in developing countries, supply shortages (especially for vaccines that target diseases or strain variants endemic primarily to developing countries), lack of donor and country financing for new vaccines, insufficient country infrastructure, lack of country awareness and political will, and competition for countrylevel health budgets. The global health community is working to close this gap and ensure that lifesaving vaccines more quickly reach those who need them most. A McKinsey study commissioned by the Gates Foundation and World Bank in 2002 recommended several steps to accelerate introduction of new vaccines into GAVI eligible countries. As a result of this recommendation, GAVI funded the Pneumococcal Vaccines Accelerated Development and Introduction Plan (PneumoADIP) in 2003 after a competitive request for proposal process selected a bid from Johns Hopkins Bloomberg School of Public Health (a similar ADIP was also formed to accelerate rotavirus vaccine). One of the goals of the PneumoADIP was to establish a secure and sustainable supply of pneumococcal vaccines for GAVI eligible countries. In developing strategies to achieve this goal, PneumoADIP evaluated the input of industry, countries and donors. One of these stakeholders, industry, indicated to McKinsey & Co. that the single biggest obstacle to successful partnership with public sector was absence of a credible strategic demand forecast. Consequently, PneumoADIP made development of a credible strategic demand forecast a high priority beginning in 2003. 5 Created for products early in their life cycle (e.g., primarily in development), strategic demand forecasts typically predict demand over a 5 to 20 year timeframe. PneumoADIP’s strategic demand forecast was the first long-term demand forecast for a global health product that leveraged industry forecasting best practices and employed a robust process for vetting the forecasting methodology, assumptions, and results. Some of these best practices include evaluating and documenting all assumptions, analyzing trends in adoption and looking at willingness to pay, The resulting strategic demand forecast was useful to a broad range of key stakeholders. Suppliers were more easily able to predict the long-term demand for their vaccines, allowing them to expand or build manufacturing capacity and set prices that deliver a neutral or better return on their research, development, and manufacturing investments. Beyond supporting supplier decision-making, strategic demand forecasts allow donors and countries to determine their potential long-term financial commitments. Prior to the PneumoADIP strategic demand forecast, supply chain forecasts spanning 1 to 3 years were primarily used by the global health community to forecast demand for global health products. The categories of expected, likely and possible demand, developed by the Procurement Reference Group (PRG), provided a wide range of potential demand scenarios. While potentially useful for vaccine procurement decisions, these forecast scenarios did not provide the longer-term perspective required to support key stakeholder decision-making. Without this long-term perspective, suppliers were reluctant to make major investments to support developing country vaccine introduction. Likewise, global donors were reluctant to commit financing without an understanding of their full liability. Countries were also unwilling to commit scarce resources without the assurance of a sustainable supply at affordable prices. As global health stakeholders, suppliers, and countries prepare to introduce the first pneumococcal vaccines into developing countries in 2009, the PneumoADIP has begun to see progress towards fulfilling its mission of accelerating access and use of pneumococcal vaccines to the world’s poorest children. The time between vaccine licensure and developing country introduction is years shorter than for any previous vaccine and overall uptake is expected to be accelerated substantially. By early 2009, a new GAVI program – Accelerated Vaccine Introduction or AVI – will assume the responsibilities of PneumoADIP and take responsibility for the coordinated introduction of pneumococcal vaccines. AVI is a partnership between GAVI, the World Health Organization (WHO), UNICEF, the World Bank, and an outsourced technical advisory consortium (AVI-TAC) comprised of the Program for Appropriate Technology in Health (PATH), Johns Hopkins Bloomberg School of Public Health (JHSPH), and the US Centers for Disease Control and Prevention (CDC). The UNICEF Supply Division will continue to develop short- to medium-term supply chain focused forecasts to ensure adequate supply and affordable pricing for pneumococcal vaccines. AVI-TAC will work closely with all AVI partners to ensure the longer-term strategic demand forecast is maintained and updated, as appropriate. This paper presents the history and results of PneumoADIP’s strategic demand forecast. It is hoped that the challenges and lessons learned experienced by PneumoADIP will help the AVI –TAC properly maintain the pneumococcal vaccine strategic demand forecast to encourage investment in the appropriate level of supply capacity and to motivate the continued development of new and more advanced pneumococcal vaccines. In addition, it is hoped this history of the PneumoADIP’s forecasting work will help others to develop accurate and credible strategic demand forecasts that will ultimately lead to the accelerated introduction of other new vaccines. 6 3. Forecasting Roundtable (2003) PneumoADIP began its strategic demand forecasting effort by seeking to understand how both industry and the public sector, the two key players involved in vaccine supply and procurement for developing countries, viewed demand forecasting. PneumoADIP convened a roundtable that brought together experts from industry and global health stakeholders, including other public-private partnerships (complete participant list provided in Appendix A). This meeting produced a number of important insights. GAVI Alliance World Bank WHO/HQ Bill & Melinda Gates Foundation Strategic Demand Forecast "Big pharma" manufacturers Emerging market manufactuers WHO/Regional offices UNICEF USAID • Strategic demand forecasts serve a different purpose than supply chain forecasts. Created for products that are typically five or more years from launch, strategic demand forecast’s represent demand over the long-term, usually 5 to 20 or more years into the future. This timeframe is critical for evaluating the risk and return associated with large vaccine-related investments such as manufacturing capacity and donor financing. Supply chain forecasts, typically spanning 18 months to 3 years, are primarily used by the global health community for vaccine procurement planning and by industry for supply chain management and short-term financial forecasting. Product need is different than product demand. The public sector and industry viewed product need and product demand differently, and it became evident that standardized definitions of these terms were necessary. The public sector used the term ‘product demand’ to mean the total number of doses required to vaccinate an entire target population. But in industry, product demand includes only those who would actually use the product, based on their willingness and ability to adopt. As a result, public sector estimates often overstated demand, leading industry to believe that the market for a given product was larger than in reality. This misalignment had the potential to cause significant tension between industry and the public sector. A transparent methodology is essential for forecast credibility and usability. The chosen strategic demand forecasting methodology must be vetted in advance with industry and global health stakeholders to confirm its credibility and ensure buy-in to the results. • • 7 • Underlying data, not just top-line conclusions, need to be shared among all industry and global health stakeholders. Similar to the forecasting methodology, all underlying demand forecast data and assumptions must be vetted with industry and global health stakeholders to allow each stakeholder to evaluate the forecast and compare it to their own forecasts, when applicable. Where possible, underlying data should come from credible public sources Better country-level data will increase accuracy in the forecasting process. One of the greatest challenges to accurate demand forecasting is the lack of robust, reliable country-level data. Disease burden data and information regarding country willingness and ability to adopt, in particular, are two important determinants of demand that were difficult to obtain from developing countries. Because they understood the constraints associated with obtaining reliable countrylevel data, the public sector was more accepting of imperfect country-level data. Industry, on the other hand, had more difficulty accepting proxies and imprecise calculations of country-level data because they were accustomed to using historical data in stable economies to generate forecasts. Industry would benefit from > 3 year demand forecasts for developing country markets. To date, the global health sector provided industry with short-term supply chain forecasts. Industry stated they would benefit from credible, multi-year demand forecasts so they could develop more realistic manufacturing capacity strategies, which in turn would most likely ensure an adequate supply of vaccines for developing country markets. . Because GAVI-eligible countries are diverse, country-by-country forecasts would be most helpful. Given the difference in disease burden, competing health needs, distribution systems, etc., across GAVI-eligible countries, industry requested country-by-country demand forecasts in order to better incorporate unique country characteristics. • • • With the Roundtable’s insights and challenges in mind, PneumoADIP created its strategic demand forecast development plan (Figure 1). The goal was to develop the methodology, design and develop a model to support forecasting on an ongoing basis, create the initial forecast, and vet it with key stakeholders. Once refined, an official strategic demand forecast – version 1.0 – would be launched at a second convening of the Demand Forecast Roundtable participants. Following launch, demand forecast assumptions would be periodically evaluated as new information emerged. Future versions of the forecast would be released, as appropriate. Figure 1. Strategic Demand Forecast Development Process 2003 2004 2005 2006 Develop Strategic Demand Forecasting Methodology 4. Demand Forecast Create Initial Strategic Demand Forecast Vet Forecast with Key Stakeholders Continually Refine Forecast Develop Strategic Demand Forecasting Model Methodology Overview 2005 Demand Forecasting Roundtable Release of SDF v1.0 2003 Demand Forecasting Roundtable 8 As a starting point for forecasting, it is important to define four key terms (Table 1). Table 1. Key Terms and Definitions Term Vaccine Need Definition The number of doses that would be required to meet the needs of the entire target population if they received a full course of treatment. Vaccine need (doses) = Vaccine need (population subjects) x Doses/Tx Potential Vaccine Market (PVM) The number of subjects expected to have access to the vaccine given country policies and infrastructure. PVM (doses) = Vaccine need (doses) x Vaccine coverage rate PVM (revenue) = Vaccine need (doses) x Vaccine coverage rate x USD/dose PVM adjusted for vaccine availability and country willingness to adopt. This demand estimate assumes no supply or financing constraints. This is the demand we hope to achieve. Potential Demand (Strategic Demand Forecast) Forecasted Demand Potential demand adjusted for available supply, global donor financing programs, country product preference, and country willingness to pay (a measure of its ability to sustain program given changes in vaccine pricing and available financing). Vaccine Need, the amount of vaccine required if the total relevant population within a country was vaccinated with the required number of doses, provided the foundation for the strategic demand forecast methodology. This upper bound estimate of demand ensures suppliers do not overproduce beyond the need for the vaccine. Recognizing that the entire relevant population is unlikely to have access to immunization services, vaccine need is adjusted by the representative vaccine coverage rate – DTP3 as the proxy for pneumococcal vaccines – to establish the Potential Vaccine Market for the vaccine. The potential vaccine market forecast is independent of vaccine availability, financing, and supply in that it conveys what the market size would be if all countries adopt a particular vaccine and achieve their forecasted vaccine coverage rates. The potential vaccine market forecast provides suppliers with a more realistic estimate of the vaccine market. The strategic demand forecast takes potential vaccine market a step further to determine the Potential Demand. It takes into account the earliest time to adoption for each country as well as the country’s time to peak coverage once the vaccine is adopted. Figure 2 illustrates this relationship. 9 Figure 2. Earliest Time to Adoption and Time to Peak Coverage Given Vaccine Availability By definition, potential demand assumes no financing or supply constraints and is, therefore, the demand we wish to achieve. Potential demand is the ultimate ambition of the strategic efforts to accelerate vaccine introduction. Underlying the strategic demand forecast is a set of assumptions which includes the following. Should any of these assumptions change so would the strategic demand forecast: • • • • • • • Pneumococcal vaccines remain a GAVI priority Availability of GAVI & donor funding throughout the forecast period GAVI application process remains in place Co-pay is not increased above willingness to pay Countries receive adequate technical support and apply for GAVI funding There is adequate supply There are no serious issues with the vaccines, e.g., safety Potential demand also provides the foundation for determining the Forecasted Demand. Forecasted demand accounts for available supply, global donor support programs, countries’ product preferences, and country vaccination program sustainability based on projected vaccine prices and a country’s willingness to pay. Willingness to pay is assessed as a country’s maximum acceptable vaccine price (MAVP). Assuming the MAVP is equal to or greater than the projected vaccine price, a country will be deemed able to sustain the vaccination program once global donor financing is no longer available. If a country is deemed unable to sustain its vaccination program (MAVP < projected vaccine price), the country will not be included in forecasted demand. Figure 3 illustrates the relationship between the vaccine market price and MAVP. 10 Figure 3. Country Vaccination Program Sustainability Assessment PneumoADIP used forecasted demand to support other program objectives, such as determining the forecasted demand impact on supplier investment returns and global donor financing policies. However, the focus of this historical review is the strategic demand forecast. PneumoADIP’s initial strategic demand forecast was the culmination of a lengthy process to develop the methodology and collect the necessary information to ensure a transparent and credible forecast. Recognizing that the information underlying the forecast would change over time as more data and information became available, PneumoADIP requested Applied Strategies develop a user-friendly tool to enable ongoing forecasting and analysis. The resulting tool, Cennium Forecaster™ for Birth Cohort Vaccines in Developing Countries, is a software application that enables users to easily modify forecast assumptions and view the results in real-time. 5. Demand Forecast v1.0 (2005-2006) With the strategic demand forecast methodology defined, PneumoADIP focused its efforts on developing transparent and credible forecast input assumptions and results. PneumoADIP, with the support of Applied Strategies, designed a collaborative process to engage key stakeholders, understand their needs, gain input into the forecast, and provide feedback on the completed forecast. Key stakeholders involved in the process, and their role in providing information or in the vetting process, is detailed in Appendix B. A. Need and Potential Market Size 11 PneumoADIP compared estimates for vaccine need and potential vaccine market size in high-, middle-, and low-income countries to demonstrate to suppliers that GAVI-eligible countries represent a large potential market. For pneumococcal vaccines, the vaccine need across all high-, middle-, and low-income countries was forecasted at approximately 420 million doses in 2010, representing full vaccination of 136 million births worldwide. The potential global vaccine market in 2010 was forecasted to be just over 350 million doses. This relationship between global vaccine need and potential vaccine market is provided in Figure 4. Figure 4. Global Vaccine Need and Potential Vaccine Market 450 400 350 300 250 200 150 100 50 0 Total Vaccine Need Total Potential Vaccine Market GAVI-eligible countries were expected to account for nearly 180 million of these doses, or just over 50% of the potential vaccine market. Figure 5 provides the vaccine need and potential vaccine market forecast for GAVI-eligible countries only. 12 Figure 5. GAVI-Eligible Country Vaccine Need and Potential Vaccine Market 450 400 350 300 250 200 150 100 50 0 Total Vaccine Need Total Potential Vaccine Market Applying expected public and private market vaccine price ranges to each of these market segments resulted in a potential global vaccine market value of approximately $5 billion, with high-, middle-, and low-income markets accounting for $2.4 billion (46%), $1.7 billion (33%), and $1.1 billion (21%), respectively. Not surprisingly, high-income markets represent 12% of potential doses but 46% of the market value, while low-income countries represent 51% of doses but only 21% of the market value. Although this dichotomy does help explain the historical delay in lowincome country adoption of global vaccines, a market value in excess of $1 billion was believed to provide suppliers adequate incentive to accelerate the introduction of these vaccines in the developing world. Table 2 summarizes these analysis results. 13 Table 2. Potential Vaccine Market Analysis Summary Vaccine Market Summary For the Year: 2010 Public Vaccine Market Dosing Regimen Potential Public Vaccine Market (M doses) Estimated Public Vaccine Price ($/dose) Potential Public Vaccine Market ($ M) Low Income Countries 3 160 $5.00 $802 Middle Income Countries 3 92 $10.00 $918 Middle Income Countries 3 39 $20.00 $788 High Income Countries 4 4 $50.00 $215 Total -257 -$1,936 Private Vaccine Market Dosing Regimen Potential Private Vaccine Market (M doses) Estimated Private Vaccine Price ($/dose) Potential Private Vaccine Market ($ M) Total Vaccine Market (M Doses) Total Vaccine Market ($ M) Low Income Countries 3 18 $15.00 $267 High Income Countries 4 39 $60.00 $2,320 Total -96 -$3,375 178 $1,070 131 $1,706 43 $2,535 352 $5,310 To support this potential vaccine market analysis, PneumoADIP made several assumptions for low-, middle-, and high-income country characteristics, including the: • • • Percentage of population receiving healthcare through public and private health systems Vaccine coverage rates in the public and private health systems Pneumococcal vaccine prices in public and private markets. Educated guess placeholders were used to estimate the percentage of the population receiving healthcare from the public and private sectors because these data were not available for lowincome countries and research was ongoing for middle- and high-income markets. In low-income countries, PneumoADIP assumed that 90 percent of vaccines were given through the public health system and 10 percent were given through private healthcare, reflecting the assumption that most individuals in low-income countries could not afford private healthcare. For middleincome countries, PneumoADIP assumed the breakdown was more even, with about half the population receiving vaccines through public health systems and half through private sources. Finally, in high-income countries, PneumoADIP assumed the majority of the population could afford private healthcare and estimated 90 percent of the population would receive vaccines 5 through private channels and 10 percent through public health . 5 After later consultation with experts, PneumoADIP learned a substantial percentage of the population in high-income countries receives vaccines through the public health system and adjusted the forecast to reflect 50% public and 50% private. 14 In addition, because data on vaccine coverage rates in the private health system were not available, the forecast assumed public and private vaccine coverage rates were equivalent. The forecast also assumed a complete course of treatment in low- and middle-income countries would require only three doses. For low-income countries, PneumoADIP projected the price per dose would be $5 in public markets and $15 in private markets. Middle-income country prices would be $10 a dose in public markets and $20 a dose in private markets. Finally, high-income countries would use four doses per course with each dose costing $54 in the public market and $65 in the private market. As Table 2 indicates, the pneumococcal vaccine market in low-income countries is substantial – $1.1 billion. This comparison between low-, middle-, and high-income countries helped vaccine manufacturers realize that low-income country markets could indeed be an attractive investment. B. Demand PneumoADIP forecasted potential demand by modifying its estimate of potential vaccine market to account for vaccine availability and country willingness to adopt the vaccine. Three parameters enable development of the potential demand forecast: 1st vaccine availability: Estimated year of first vaccine availability, assessed as year of licensure for high- and middle-income countries and as year of WHO prequalification for GAVI-eligible or low income countries. Earliest time to adoption: The minimum number of years after first vaccine availability a country will require before adopting a given vaccine. The delay from availability could reflect country economic conditions, competing healthcare needs, attitudes about immunization, political or social issues, etc. Time to peak coverage: The number of years a country will take to reach its peak vaccination coverage rate. i. 1st Vaccine Availability To determine the year of first vaccine availability, PneumoADIP and Applied Strategies conducted a vaccine landscape analysis. One component of the vaccine landscape analysis identified pneumococcal vaccines in the clinical development pipeline. Figure 6 summarizes the vaccine development pipeline as it existed in 2005. 15 Figure 6. Pneumococcal Vaccine Development Pipeline (2005) At the time of the landscape analysis in 2005, Wyeth’s 7-valent Prevenar vaccine was already on the market, but at that time, it was not recommended for use in developing countries and it had not yet been WHO prequalified. GSK’s 10-valent and Wyeth’s 13-valent vaccines were considered appropriate vaccines for developing country markets and both were in Phase III clinical development. Assuming Phase III success, vaccine licensure for both the 10- and 13valent vaccine was expected in 2009 and WHO prequalification in 2010. Therefore, demand is not expected prior to 2010. ii. Earliest Time to Adoption One of the greatest challenges for PneumoADIP in determining potential demand was developing the country adoption forecast. Retrospective analyses of vaccine adoption clearly demonstrated the delay between industrialized and developing country adoption. Figure 7 summarizes the adoption lag times associated with Hepatitis B and Hib vaccines. 16 Figure 7. HepB & Hib Introduction Timeline To better understand the principles of new product adoption, PneumoADIP worked with Paul Wiefels, author of The Chasm Companion and a product adoption expert in the information technology industry, to better understand product adoption cycles and obstacles. Wiefels conducted two one-day workshops on the premise of his book – that discontinuous innovations (new products never before seen in the market) create a chasm that delays or stops introduction of a product. In the workshop, Wiefels posited that new vaccine products for developing countries are a discontinuous innovation. New vaccine products often originate in industrialized countries, which have the resources to adopt the vaccines even if prices are high or the number of deaths averted is relatively small. These Early Adopters are willing to pay a higher price to be among the first to use a given vaccine because they recognize any type of health benefit, however large or small. The second group to adopt new products is the Mid Adopters. Mid adopters see the potential benefit of a given vaccine, but they want more information, preferably from trusted partners, as well as lower prices. As a result, there is a time lag – an adoption chasm – between the early and mid adopters. Historically, there has been an adoption lag between industrialized and early adopter developing countries of 15 to 20 years. However, once the early developing country adopters introduced a vaccine, there was a further delay before the mid developing country adopters began introducing the same vaccine. Late Adopters then follow the mid adopters. These late adopters are not necessarily resistant to vaccine introduction; they may just lack the infrastructure or financing to successfully introduce a new vaccine. Wiefels’ paradigm helped PneumoADIP understand they actually faced a twofold challenge. Not only did PneumoADIP have to decrease the time between new vaccine introduction in industrialized and developing countries, but they also had to bridge the gap between early and mid adopter developing countries to truly accelerate the introduction of pneumococcal vaccines in GAVI-eligible countries. 17 To assess developing country adoption, PneumoADIP identified key country characteristics it believed would drive the country adoption forecast. These characteristics were summarized by a country’s need for a vaccine, and their willingness and ability to adopt a vaccine. To assess a country’s need for the vaccine, PneumoADIP gathered country level data on acute respiratory infection (ARI) deaths per year as a proxy for disease burden. It also assessed whether competing disease, such as malaria, would have a high, medium, or low impact on the pneumococcal vaccine adoption decision. To assess a country’s willingness to adopt the vaccine, PneumoADIP documented whether a country currently had disease surveillance networks in place and whether they had already adopted other recent vaccines, such as HepB and Hib. Finally, PneumoADIP assessed a country’s ability to adopt a pneumococcal vaccine based on their DTP3 coverage rate, as a proxy for infrastructure health, and GNI per capita, as a proxy for their ability to pay for a new vaccine. Based on results of the country characteristic assessment, PneumoADIP categorized countries as: • • • Early adopters: Countries that would adopt within the first 5 years of vaccine availability Mid adopters: Countries that would adopt within the first 10 years of vaccine availability Late adopters: Countries that would adopt 10 years or more after vaccine availability The initial PneumoADIP country adoption categorization was then vetted with external experts, including 26 international experts with country-level immunization program expertise representing WHO, UNICEF, USAID, and Expanded Program on Immunization (EPI) managers. These experts provided additional country-specific data, as well as anecdotal evidence, to explain expected country adoption behavior and help PneumoADIP refine its country adoption categorization. Although PneumoADIP also considered country consultations to further refine the adoption categorization, in the end, they were avoided based on a concern for the time and cost involved for a perceived small gain in long-term forecast resolution. PneumoADIP’s assessment resulted in 26 early adopter countries expected to adopt within six years of vaccine availability to GAVI-eligible countries. Twenty-five countries were expected to be mid adopters (2016-2020) and 21 countries late adopters (2021-2025). The resulting country segmentation and each country’s earliest time to adoption are summarized in Table 3. 18 Table 3. Country Categorization and Earliest Time to Adoption Assessment Figure 8 summarizes PneumoADIP’s country adoption forecast graphically. Figure 8. Country Adoption Forecast 19 iii. Time to Peak Coverage After determining the earliest time to adoption, PneumoADIP estimated the time it would take countries to reach peak vaccination coverage. These estimates were based on input from WHO and UNICEF. Based on country-level health policies, infrastructure, and historical data from the introduction of HepB and Hib vaccines, early adopter countries were assumed to reach peak coverage within two years. Mid adopters generally demonstrated relatively weaker infrastructures and were estimated to achieve peak coverage within three years. Late adopters were assumed to have exceptionally underdeveloped infrastructures and were estimated to take four years to reach peak coverage. iv. Results The potential demand forecast resulting from these assessments for first vaccine availability, earliest time to vaccine adoption, and time to peak coverage is provided in Figure 9. By 2015, potential vaccine demand was forecasted at approximately 50 million doses, increasing to nearly 160 million doses by 2020, and peaking at just over 200 million doses by 2030. This forecast does not include potential vaccine wastage or doses procured for use as buffer stock. Assuming pneumococcal vaccines have the potential to prevent 7 deaths per 1000 children 6 vaccinated , approximately 5.3 million deaths could be averted between 2010 and 2030. Figure 10 provides the potential annual and cumulative deaths averted as a result of pneumococcal vaccine immunization. Figure 9. Potential Demand Forecast 220 200 180 160 140 120 100 80 60 40 20 0 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 6 Cutts, FT, et al. “Efficacy of nine-valent pneumococcal conjugate vaccine against pneumonia and invasive pneumococcal disease in The Gambia: randomized, double-blind, placebo-controlled trial.” The Lancet, 2005: 1139 – 1146. Demand (M Doses) 20 Figure 10. Potential Annual and Cumulative Deaths Averted 450 Annual Deaths Averted (1000s) 6,000 5,000 Cumulative Deaths Averted (1000s) 400 350 300 250 3,000 200 150 100 1,000 50 0 2010 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2011 4,000 2,000 0 Annual Deaths Averted (1000s) Cumulative Deaths Averted (1000s) C. Demand Forecast Impact PneumoADIP launched its initial strategic demand forecast at a 2005 Demand Forecasting Roundtable that included available participants from the 2003 Roundtable plus others (Appendix A). During this meeting, industry representatives endorsed the strategic demand forecast methodology and expressed cautious confidence in PneumoADIP’s v1.0 forecast. Industry confirmed PneumoADIP met its request for a transparent, industry-caliber forecast with wellvetted input assumptions. However, industry expressed their concern that the forecast be continually updated to reflect changes in assumptions necessitated by the complex and dynamic developing country market. Industry also indicated the next forecasting challenge would be to account for the availability of vaccines for other high burden diseases. Following the v1.0 launch, PneumoADIP presented the strategic demand forecast results to a myriad of groups, including other vaccine manufacturers, the Developing Country Vaccine Manufacturers Network, the International Symposium of Pneumococci and Pneumococcal Diseases (ISPPD), global health organizations, global donors, NGOs, and PDPs. The forecast was widely recognized as the new standard for strategic demand forecasting in the global health sector. PneumoADIP’s strategic demand forecast methodology and input assessment process was used as a model for other strategic demand forecasting efforts. Most notably, the Rotavirus Vaccine Program (RVP) at PATH developed its strategic demand forecast for rotavirus vaccines, successfully leveraging PneumoADIP’s methodology and Cennium Forecaster model to make the case for rotavirus vaccine introduction. PneumoADIP’s strategic demand forecast also supported the GAVI Investment Case for pneumococcal vaccines. For the Investment Case, PneumoADIP leveraged the potential demand forecast to demonstrate to GAVI the impact different supply and financing constraints would have on demand, as well as deaths averted. This helped GAVI understand the liability it would face 21 given different financing terms and the risk associated with inadequate or oversupply scenarios. As a result of this Investment Case, the GAVI Board approved country financing support for pneumococcal vaccines and increased its investment in the PneumoADIP to support additional vaccine readiness activities. In addition to helping GAVI understand the implications of pneumococcal vaccine introduction support, PneumoADIP’s strategic demand forecast helped industry make viable business cases within their own organizations for the continued investment in pneumococcal vaccine development and supply. The process and collaboration of stakeholders in the public sector helped give added confidence to the reliability of the forecast. The strategic demand forecast opened the door for meaningful discussions between the global health community and suppliers on potential pricing, capacity strategy, and participation in developing markets. Regular visits to industry and discussion regarding details of demand generation activities have led to a vision of shared success where industry not only has expectations of public sector, but plays a more active role in generating data in developing countries, participating in educational and advocacy activities and providing support to address issues that impact demand uptake in these countries. PneumoADIP’s strategic demand forecast also formed the basis for the Advance Market Commitment financial implications and risk model (AMC-FIRM). The AMC is an innovative financing mechanism to support procurement of vaccines for GAVI-eligible countries. The AMCFIRM analysis of pneumococcal and other vaccines supported the February 2007 decision to fund a pilot $1.5 billion pneumococcal vaccine AMC. The strategic demand forecast was an important first step in bringing long-term demand forecasting to the global health arena. However, like any new innovation, the strategic demand forecast v1.0 had both strengths and weaknesses. Its strengths included: Successful inclusion of industry and global health stakeholders: Previous demand forecasts for global health products were criticized for their lack of methodology and assumption transparency, and they were developed without industry and other key stakeholder input. As a result, the forecasts were often inaccurate, creating skepticism and slow responses from industry. Credible market potential forecasts for multinational and developing country suppliers: The vaccine need, potential vaccine market, and potential demand forecasts helped shape industry’s perception of developing country markets. Multinationals recognized that viable markets can exist in developing countries, especially for products that can be sold in highincome countries at higher prices to recover the significant investment in vaccine R&D. The forecast enabled companies to make a business case for shareholders to invest in products for developing countries. The strategic demand forecast also encouraged developing country manufacturers, who were further behind in the development process, to continue to invest in pneumococcal vaccine development. Availability of a user-friendly model capable of providing real-time analyses of alternative assumption scenarios. The Cennium Forecaster model allowed users to alter their underlying assumptions and immediately see the impact on potential demand. The weaknesses associated with strategic demand forecast v1.0 included: 22 The forecast is only as good as its assumptions. Better data are needed at the country level to support more credible forecasting. Many variables supported the country adoption forecast, several of which were assessed subjectively. Additional country-level research, including conjoint analyses with decision maker input, could provide the basis for a more realistic assessment of country adoption behavior. The forecast does not account for the presence of multiple new vaccines to address other diseases. New vaccines are being developed more rapidly, and countries have many options for addressing their healthcare needs. Given the constraints on country financing and infrastructure, introduction of multiple interventions at the same time may be unrealistic. The potential for alternative or multiple new vaccine adoptions and competing health interventions will ultimately influence demand for pneumococcal vaccines. Vaccine uptake may not occur in a linear fashion. The assumption that vaccine uptake occurs in a linear fashion once the vaccine is introduced is a likely oversimplification. In the larger countries, adoption may vary from state to state. In more challenged countries, adoption may focus on high access areas until conditions allow for broader implementation. Potential demand did not account for wastage or buffer stock. The strategic demand forecast focused on the doses required to meet actual immunization needs. Because wastage is a function of the vaccine itself, Cennium Forecaster only accounts for wastage when calculating forecasted demand. Buffer stock was considered a supply chain forecast issue and ignored for potential demand forecasting purposes. 6. Strategic Demand Forecast v2.0 (2006-2008) As the global health community moved closer to pneumococcal vaccine introduction, policy environment and other key assumption changes motivated a 2006 update to the v1.0 strategic demand forecast. In 2006, WHO’s Strategic Advisory Group of Experts (SAGE) recommended 7-valent pneumococcal conjugate vaccine for use in select developing country markets. To understand the impact of this recommendation, GAVI sent letters in 2007 to GAVI-eligible country health ministers soliciting their nonbinding expression of interest in adopting pneumococcal vaccines as early as 2008. In the letter, GAVI explained that Prevenar would be available immediately, but cautioned that Prevnar supply would be constrained after 2010 when the 10- and 13-valent vaccines were expected to be licensed. Thirty-four expression of interest letters were subsequently received by GAVI, which provided invaluable information for updating the pneumococcal vaccine strategic demand forecast. Once it became clear that an updated strategic demand forecast was required, PneumoADIP considered all the information gathered since the v1.0 launch and also confirmed that the underlying population and coverage data had been updated by the UN and WHO, respectively. To understand how each data or assumption change impacted the v1.0 forecast, PneumoADIP conducted a step-wise update of its v1.0 forecast. PneumoADIP first explored the impact of the updated UN population forecast. Incorporating the latest UN population database forecast increased total GAVI-eligible country births by approximately 1.7%. 23 PneumoADIP then determined the impact of the updated WHO ICE-T DTP3 coverage rate forecast. Incorporating the latest DTP3 forecast increased the total birth population coverage by approximately 4 million to 6 million births per year. Taken together, these changes created an 8% steady-state increase in potential demand, or approximately 15 million additional vaccine doses per year. The 2010 to 2015 cumulative demand was 50% greater than v1.0, resulting in an additional 35 million vaccine doses required during this sensitive time period. Figure 11 summarizes the combined effect of changes to these underlying forecasts. Figure 11. Cumulative Impact of Population and DTP3 Coverage Rate Forecasts 2005 DF v1.0 (1000s) Change (1000s) Percent Change ‘06 ‘07 ‘08 ‘09 ‘10 ‘11 ‘12 ‘13 ‘14 ‘15 ‘16 ‘17 ‘18 ‘19 ‘20 ‘21 ‘22 ‘23 ‘24 ‘25 ‘26 ‘27 ‘28 ‘29 2030 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1.2 0.9 73 5.1 3.6 71 10 4.5 44 22 7.1 32 34. 12 35 46 6.3 13 52 5.2 10 73 11 15 97 20 21 121 152 166 176 186 188 193 195 198 200 201 201 29 24 18 12 18 11 23 13 15 8 16 8 15 8 15 8 16 8 17 9 16 8 15 7 Once the strategic demand forecast was updated to reflect changes in population data and DTP3 coverage rates, PneumoADIP estimated the impact of the 34 expression-of-interest letters on the country adoption forecast. The majority of these countries (22) indicated a desire to introduce pneumococcal vaccines by 2010. The remaining 12 countries expressed interest without a specified date of introduction. Only 17 of these countries had previously been categorized as early adopter countries – the others having been categorized as mid-adopters (13) or late-adopters (4). Table 4 summarizes the information gleaned from the GAVI expression of interest letters. 24 Table 4. Country Expression of Interest Letter Information Summary Although a thorough country consultation or in-depth survey would be required to truly understand the greater than expected interest in pneumococcal vaccines, several explanations emerged. PneumoADIP had conducted significant media and advocacy campaigns to increase country awareness of their pneumococcal disease burden. The launch of the pneumococcal vaccine pilot AMC, with $1.5 billion of committed funds to support vaccine procurement, most likely sent a strong signal that the global community was serious about supporting pneumococcal vaccine introduction. In addition, the availability of Prevenar might have convinced skeptical country decision-makers that pneumococcal vaccines could be affordable. Finally, political and economic conditions in some countries might have changed, leading to new health priorities and interest in pneumococcal vaccines. Nine countries who were projected to be early adopters did not submit expressions of interest letters. These countries were left as likely early adopters except Eritrea, which was reclassified as a late adopter due to its change in GAVI status to a fragile state. The earliest time to adoption for the mid adopter countries were either left unchanged or advanced one year. Adoption time was also advanced one year for all late adopter countries, which resulted in Kiribati being moved to the mid adopter category. 25 As PneumoADIP finalized the country adoption forecast assumptions, countries were also adopting or preparing to adopt Hib vaccines. Assuming that countries could not adopt both vaccines simultaneously, PneumoADIP pushed pneumococcal vaccine adoption back by three years for countries that were planning to simultaneously adopt the Hib vaccine. Hib vaccines took precedence because they were already available and WHO prequalified. Table 5 summarizes the adoption forecast changes resulting from an analysis of the expression of interest letters and expected Hib vaccine adoption. Table 5. Updated Expression of Interest Country Adoption Assumptions Country Gambia Kenya Benin Mali Senegal Ghana Mongolia Nicaragua Sri Lanka Malawi Sudan Uganda Yemen Zambia Burundi Rwanda Cuba Guyana Tanzania Togo Côte d'Ivoire Ethiopia Sao Tome and Principe Solomon Islands Pakistan Madagascar Congo, DRC Honduras Indonesia Moldova Djibouti Afghanistan Timor-Leste, DR Congo WHO Region AFR AFR AFR AFR AFR AFR WPR AMR SEA AFR EMR AFR EMR AFR AFR AFR AMR AMR AFR AFR AFR AFR AFR WPR EMR AFR AFR AMR SEA EUR EMR EMR SEA AFR Demand Forecast v1.0 Adoption Category Earliest Adoption Demand Forecast v2.0 Adoption Category Earliest Adoption Change from Demand Forecast v1.0 EOI 2008; GAVI NVS Sep08; assume 1 year delay EOI 2008; conditional GAVI NVS May08; assume 1 year delay EOI 2009; but Hib in 09, therefore 3 years after Hib EOI 2009; GAVI NVS Sep08; assume 1 year delay EOI 2009; but MenA in '10, therefore 2 years after MenA EOI 2010; assume 1 year delay EOI 2010 EOI 2008 EOI no date; GAVI NVS May08; moved up 3 years EOI no date; left at 2013 EOI 2010; GAVI NVS May08; assume 5 yrs to peak given unrest in South EOI no date; left at 2013 EOI 2008; conditional GAVI NVS Feb08; assume 1 year delay EOI no date; left at 2013 EOI no date; moved up 1 year EOI 2008; GAVI NVS May08; assume 1 year delay GAVI NVS Sep08 with adoption expected 2009 EOI 2008 GAVI App in 09; assume 2 year delay EOI 2009; MenA/Hib in 09, therfore 4 years after MenA/Hib EOI 2009; but Hib in 09, therefore 3 years after Hib EOI no date; moved up 6 years EOI no date; Hib in 09, therefore 3 years after Hib EOI 2008; but Hib in 09, therefore 3 years after Hib EOI 2008; GAVI NVS Sep08; but Hib in 09, therefore 3 years after Hib EOI 2010; but Hib in 09, therefore 3 years after Hib EOI no date; but Hib in 09, therefore 3 years after Hib EOI 2009 EOI no date; moved up 6 years GAVI NVS Sep08; but Hib in 09, therefore 3 years after Hib EOI 2010; GAVI NVS May08 GAVI NVS Sep08; but Hib in 09, therefore 3 years after Hib EOI no date; moved up 10 years EOI 2009; but Hib in 09, therefore 3 years after Hib Early Mid Late 2010 2010 2011 2011 2011 2012 2012 2012 2012 2013 2013 2013 2013 2013 2014 2014 2015 2016 2016 2016 2017 2017 2017 2017 2018 2019 2020 2020 2020 2020 2022 2023 2023 2024 Early 2009 2009 2012 2010 2012 2011 2010 2008 2009 2013 2010 2013 2009 2013 2013 2009 2009 2008 2011 2013 2012 2011 2012 2012 2012 2012 2012 2009 2014 2012 2010 2012 2013 2012 After all adjustments were made, the v2.0 forecast assumptions predicted a significantly higher number of early adopters and a reduced number of mid and late adopters. A comparison of the v1.0 and v2.0 country adoption forecasts is provided in Figure 12. 26 Figure 12. Country Adoption Forecast Comparison (v1.0 versus v2.0) 16 14 # of Countries Adopting 12 10 8 6 4 2 0 Year of Adoption DF v1.0 DF v2.0 Category Early Mid Late Demand Forecast v1.0 26 25 21 Demand Forecast v2.0 42 15 15 Accounting for these changes in expected country adoption, strategic demand forecast v2.0 estimated an additional 250 million doses of pneumococcal vaccine over the v1.0 forecast would be required in the 2008 – 2015 timeframe and an additional 475 million from 2008 through 2020. Given emerging suppliers are not expected to enter the market until 2015 or beyond, significant global supplier capacity would be required to avoid shortages in the near-term. A comparison of the v1.0 and v2.0 strategic demand forecast is provided in Figure 13. Figure 13. Strategic Demand Forecast Comparison (v1.0 versus v2.0) 27 If supply can be made available, approximately 2.5 million additional deaths could be averted between 2008 and 2030 through the accelerated introduction of pneumococcal vaccines. A comparison of the v1.0 and v2.0 potential annual and cumulative deaths averted forecast is provided in Figure 14. Figure 14. Potential Annual and Cumulative Deaths Averted (v1.0 versus v2.0) 7. Strategic Demand Forecast Performance Strategic demand forecasts are intended to predict the shape of a market 5 to 20 years in the future. Because of their nature, these forecasts are uncertain and based on benchmarks, broad assumptions, and limited data. As a result, evaluations of their accuracy should only be conducted for the purpose of improving upon the methodology or assessment assumptions. Continual evaluation is critical to ensuring the next generation of strategic demand forecasts build upon the lessons learned of earlier forecasters. To develop these lessons learned, the PneumoADIP strategic demand forecasts were evaluated against currently available information. Since pneumococcal vaccines have not yet been introduced into GAVIeligible countries, there are relatively little data available. At this point in time, only vaccine availability and country adoption assumptions can be evaluated based on actual submissions for licensure and approved GAVI applications. PneumoADIP’s v1.0 forecast assumed the 7-valent pneumococcal vaccine would not be introduced in GAVI-eligible countries; therefore, adopters were not expected until 2010. The v2.0 forecast did assume 7-valent availability, but assumed vaccination would begin in 2008. In either case, a one year shift in early adopter vaccine introductions has relatively little impact on the overall accuracy of the forecast. 28 The forecast is consistent with the anticipated licensure and WHO prequalification of both the 10- and 13valent vaccines. Both versions of the forecast predicted the 10- and 13-valent vaccines would be licensed in 2009 and WHO prequalified in 2010. Given both vaccines were filed for licensure in 2008, vaccine availability assumptions are likely to hold. The accuracy of this underlying assessment is reflective of PneumoADIP’s significant investment in supplier relationship-building and vaccine pipeline monitoring and its detailed understanding of the licensure and WHO prequalification process. PneumoADIP’s country adoption forecast can be evaluated based on GAVI application data. As of December 2008, 11 country applications for pneumococcal vaccine have been approved by GAVI and 1 country has received conditional approval. Of these 12 countries, 6 had been initially designated as early adopters, 3 as mid adopters, and 3 as late adopters. An analysis of these differences highlights the ‘art, not science’ aspect of country adoption forecasting. Table 6 summarizes the results of the forecast versus actual country adoption assumptions analysis. Even though there are only limited data to enable country adoption forecast evaluation, it is clear countrylevel intelligence is critical for strategic demand forecasting. Active participation in adoption-related assessments by organizations such as WHO, PAHO, and UNICEF could greatly enhance the credibility of this key forecasting component. As seen in some of the variances, past performance is by no means an indicator of future performance and clearly, education and demand generation efforts and ensuring financing have influenced country decisions. Country decision-maker input would also be invaluable; however, this input would be significantly more challenging and costly to collect. 29 Table 6. Forecast vs. Actual Country Adoption Assumptions Actual vs. v1.0 (years) Actual vs.v2.0 (years) Country v1.0 Rationale v2.0 Rationale Actual Rationale • Strong global health stakeholder push for pneumo vaccine adoption in West & Central Africa • GAVI application approved December 2008 • Strong global health stakeholder push for pneumo vaccine adoption in West & Central Africa • GAVI application approved July 2008 • Strong in-country advocacy • GAVI application approved July 2008 • Strong global health stakeholder push for pneumo vaccine adoption in West & Central Africa • GAVI application approved December 2008 • Unexpected availability of 7valent Prevnar • GAVI application approved July 2008 • PAHO strongly advocating for pneumo vaccine adoption • GAVI application approved 2007 • PAHO strongly advocating for pneumo vaccine adoption • GAVI application approved 2007 • Strong in-country advocacy • Pneumo vaccine evaluation study planned • GAVI application approved July 2008 • On political agenda • GAVI application conditionally approved November 2008 Cameroon 2024 • Late Hib adopter 2023 • Across the board late adopter update • Across the board late adopter update • EOI submitted (no date) • Hib in 2009 – wait 3 yrs • EOI submitted for 2009 • Hib in 2009 – wait 3 yrs • EOI submitted for 2008 • Assumed 1 yr wait for 10- or 13-v • EOI submitted for 2008 • EOI submitted for 2009 • EOI submitted for 2008 • EOI submitted for 2009 • Assumed 1 yr wait for 10- or 13-v • EOI submitted for 2008 • EOI submitted for 2008 • Assumed 1 year wait for 10- or 13-v • EOI submitted for 2008 • Assumed 1 year wait for 10- or 13-v 2010 (10-v or 13-v) -14 -13 Central African Republic 2024 • Late Hib adopter • GAVIdesignated fragile state • Late Hib adopter • GAVIdesignated fragile state • Late Hib adopter • GAVIdesignated fragile state • Conducted pneumo vaccine clinical trial • Stable government with strong interest in vaccination • Political instability • Netspear site • Pneumo disease burden research 2023 2010 (10-v or 13-v) -14 -13 Democratic Republic of Congo 2020 2012 2010 (10-v or 13-v) -10 -2 Republic of the Congo 2024 2012 2010 (10-v or 13-v) -14 -12 The Gambia 2010 2009 2009 (Prevna r) -1 0 Guyana 2016 2008 2010 (10-v or 13-v) 2010 (10-v or 13-v) 2009 (7valent) -6 +2 Honduras 2020 2009 -10 +1 Kenya 2010 2009 -1 0 Mali 2011 • Early Hib adopter • Stable government, demonstrated interest in adopting pnemo vaccine • Strong infrastructure and political will but innovation pile-up delayed to 2014 • Expressed strong interest in adopting pneumo, but not ready to adopt in 2010 2010 2010 (10-v or 13-v) -1 0 Nicaragua 2012 2008 2009 (7valent) • PAHO strongly advocating for pneumo vaccine adoption • GAVI application approved 2007 • Local immunization NGO promoting pneumo vaccination • GAVI application approved July 2008 • On political agenda • Impact of disease burden awareness and education programs • GAVI application approved December 2008 -3 +1 Rwanda 2014 2009 2009 (7v) -5 0 Yemen 2013 2009 2010 (10-v or 13-v) -3 +1 30 8. Demand Forecast Ongoing Monitoring and Evaluation Eleven GAVI countries have currently been approved to receive GAVI support for pneumococcal conjugate vaccine. Rwanda and the Gambia will adopt Prevenar 7-valent vaccine, receiving the product as a Wyeth donation. Three GAVI countries in the Americas region planned on adopting Prevenar in 2008, but introduction was delayed pending funding agreements between PAHO and GAVI. The other six countries will wait for the 10- or 13-valent product to achieve better serotype coverage and to avoid waste disposal issues associated with Prevenar’s single dose syringe packaging. Although adoption is underway, the longer-term strategic demand forecast will still play a critical role in helping key stakeholders make informed decisions in the face of an ever-changing market environment. Updated forecasts will be needed to encourage the development of new pneumococcal vaccines, including conjugate vaccines from emerging suppliers and the next generation protein vaccines. Ongoing monitoring and evaluation of the strategic demand forecast is also important for managing stakeholder expectations. If the underlying assumptions supporting the forecast do not pan out as expected, the implications could be significant. Semi-annual forecast evaluations and updates with implication analyses are highly recommended to support the most efficient adoption and uptake of this lifesaving vaccine. On a larger scale, as multiple vaccines become available simultaneously, countries will face tough decisions about which vaccines to introduce and when. Their decisions will need to take into account the trade-offs between disease burden, cost effectiveness, infrastructure requirements, potential product health impact, as well as other political and cultural factors. Strategic demand forecasts will need to account for the potential impact multiple new vaccine options will have on the decision to introduce a specific vaccine. Many stakeholders will need to provide input and guidance on how best to account for multiple vaccine introductions in strategic demand forecasting. 31 Appendix A: 2003 & 2005 Demand Forecasting Roundtable Participants 2003 Roundtable Name Dr. Oya Afsar Gursh Bindra Alan Brooks Dr. Maria Deloria-Knoll John Fitzsimmons Shawn Gilchrist Chris Gingerich Steve Jarrett Dr. Subash Kapre Hans Kvist Steve Landry Dr. Orin Levine Dr. Yvette Madrid Thomas McDowell Susan McKinney Eunice Miranda Farzana Muhib Thomas Netzer Evan Simpson Dr. John Wecker Affiliation WHO Immunization, Vaccines, and Biologicals Aventis PATH PneumoADIP PAHO Aventis Bill and Melinda Gates Foundation UNICEF Developing Country Vaccine Manufacturers Network PneumoADIP GAVI Vaccine Fund PneumoADIP Rotavirus Vaccine Program Wyeth US Agency for International Development GlaxoSmithKline PneumoADIP Merck Rotavirus Vaccine Program Rotavirus Vaccine Program Others who were invited, but were unable to attend: Name Amie Batson Violaine Mitchell Raj Shah Affiliation World Bank GAVI Financing Task Force Bill and Melinda Gates Foundation 32 Appendix A: 2003 & 2005 Demand Forecasting Roundtable Participants (cont’d) 2005 Roundtable Name Deborah Atherly Luigi Bonfatti Stephen Brooke Laura Efros Elaine Esber John Fitzsimmons Mary Gadek Gian Ghandi Rebekah Heinzen Walt Jones Miloud Kaddar Maria Knoll Lynda Kulp Nand Kumar Steve Landry Ruth Levine Orin Levine Jim Maleckar Carol Marzetta Susan McKinney Michelle Moncrieffe-Forman Farzana Muhib Angeline Nanni Lois Privor-Dumm Eva Roca Craig Schaffer Wendy Taylor Ross Underwood Walter Vandermissen Earl Wall Hugh Waters James Watt Alastair West Jessica Wolf Sandra Wrobel Affiliation PATH GSK PATH - HPV Merck Merck PAHO Sanofi IAVI PneumoADIP Hib Initiative WHO PneumoADIP Wyeth Wyeth Bill and Melinda Gates Foundation CGD PneumoADIP Sanofi Applied Strategies US Agency for International Development PneumoADIP PneumoADIP PneumoADIP Hib Initiative IAVI Applied Strategies Bioventures for Global Health Merck GSK PneumoADIP Johns Hopkins University Hib Initiative World Bank CGD Applied Strategies 33 Appendix B: Key Forecast Variables with Sources and Vetting Groups Variable Target population Definition Birth cohorts of the 72 countries eligible for GAVI financial support, defined as countries with gross 5 national incomes less than $1000 per capita. Source/vetting groups UN 2004 Population Division, Department of Economic and Social Affairs, World Population country birth cohort data used for all calculations total population from stock indicators total population both sexes, for 1974–2000 estimates for 2001–2010 medium variant births from supplementary tabulation: Interpolated demographic indicators 1974–2010,births, medium variant; WHO, UNICEF PneumoADIP Supply Strategy Working Group; Industry Introduction year First year the vaccine is available to GAVI-eligible countries. Availability to GAVI-eligible countries assumes vaccine supplier has received pre-qualification from WHO. Including 7-valent vaccine available in 2008, 10-valent vaccine expected in 2009, and 13-valent vaccine expected in 2010, additional suppliers’ potential availability 2015 through 2017. Estimates of DTP3 coverage for each country and each year. As DTP3 is given to infants on the same schedule as pneumococcal vaccine this was selected as a good indicator. The number of years the country will take to reach its peak vaccination coverage rate. It is assumed that introduction of pneumococcal vaccine into a national program for early, mid, and late adopters would require 2, 3 or 4 years, respectively, to reach peak levels (peak coverage was defined as DTP3 coverage in that year). The build-up to peak was assumed to be evenly spaced over time. For example, countries requiring 2 years would reach 50% of peak coverage in the first year and full coverage the second year. Number of doses of pneumococcal vaccine administered to each infant. Based on data from efficacy and immunogenicity trials, pneumococcal vaccination was assumed to require three doses given in infancy, on the same schedule as DTP vaccine. Vaccination coverage rate WHO: Immunization Coverage Estimates and Trajectory database (WHO ICE-T 2005) Years to peak adoption WHO EPI regional experts input and review Doses per course Industry data from efficacy and immunogenicity trials 34 Wastage rate Percentage of vaccine that is likely to be rendered unusable as a result of spoilage, breakage in transit, being part of an unused open package, at the country level. Estimated at 10%. The minimum number of years the country will take before it adopts the available vaccine. Based on both quantitative and qualitative analysis countries were allocated to one of three segments (early, mid, or late adopters) and assigned an “earliest time to adoption”. An estimate of whether a country wants to adopt a vaccine. Disease burden was assessed by: 1) the burden of child pneumonia deaths; 2) the ability to measure pneumococcal disease within the country (e.g., disease surveillance in place); and 3) the presence or absence of competing diseases (e.g., HIV/AIDS, malaria, meningococcal disease); and 4) the country’s history of adopting new vaccines (HepB and Hib). Countries were segmented by high pneumonia deaths (>10,000/yr); medium pneumonia deaths (1,000–9,999/yr); and low pneumonia deaths (<1,000/yr). The ability to measure disease was either “yes” or “no” based on the presence of disease surveillance. A significant competing disease was considered a disease that was documented to have a significant impact on childhood survival and was assessed as either high (more than 1 significant competing disease), medium (1 significant competing disease), or low (no significant competing disease). An estimate of whether a country was able to adopt a new vaccine. Estimated as a function of each country’s vaccination infrastructure, economic strength and stability, and the ability of a country to sustain vaccination after donor funding ends. The vaccination coverage rate of DTP3 was used as a proxy of country vaccination infrastructure with countries segmented by high coverage rates (≥80%), medium coverage rates (66–79%), and low coverage rates (<65%). Country economic strength and stability was assessed by GNI per capita with countries segmented by high GNI (≥$700), medium GNI ($401–$599), and low GNI (≤$400) and known political instability (e.g., civil unrest or war). The vaccine price each country has the ability to pay after the donor funding ends. The model includes a “sustainability” assumption that assumes a country would not adopt a new pneumococcal vaccine unless it had the ability to WHO vaccine wastage rate estimates for liquid formulation in 1–2 dose vials Earliest time to adoption PneumoADIP estimates, WHO regional experts, international experts from GAVI Alliance, USAID, UNICEF Willingness to adopt Interview of 26 international experts from WHO regional offices, GAVI Alliance, USAID, UNICEF, Gates Foundation WHO pneumonia mortality data Ability to adopt Interview of 26 international experts from WHO regional offices, GAVI Alliance, USAID, UNICEF, Gates Foundation Maximum acceptable vaccine price PneumoADIP analysis 35 sustain procurement of the vaccine after donor funding ended. Based on the cost-effectiveness data, the high disease burden, and the consistently high vaccine impact with pneumococcal conjugate vaccines, it was assumed that all GAVI-eligible countries, along with their local partners, would be willing to accept a price of $2 per dose as a sustainable price. Country co-pay The amount of expected financing from countries and partners for vaccine purchase. For this analysis, calculations were run with two different levels of country co-pay based on the ranges set by the GAVI Alliance. These co-pay ranges are: • “Least poor countries” – co-payment equals $0.70 to $0.95 per dose • “Intermediate countries” – co-payment equals $0.20 to $0.50 per dose • “Poorest countries” – co-payment equals $0.10 to $0.25 per dose • “Fragile countries” – no co-payment required For the calculations in this analysis, estimates were run using the “low” and “high” end of the copayments. Each country was assigned a co-pay based on which of the four groups it currently belongs. Analysis was run separately for the “low” and the “high” co-payment assumptions and is kept constant between 2008 and 2015. The expected vaccine market price to GAVI per year over time. The analysis assumed a constant price to GAVI of $5 per dose until 2015. During this period it is expected that countries will make a small copayment but that the price to countries will be highly subsidized by GAVI financing. Beyond 2015, the price to countries and their local partners is assumed to be $2 per dose. The discount rate used for present value calculations of suppliers (10%), donors (5%) and countries and/or country partner donors (5%) cash flows. GAVI Alliance Vaccine price PneumoADIP analysis Supplier, donor, and country discount rates GAVI Fund, industry 36