Modeling can provide key input into public health decisions to use, or not use, new health technologies in the developing world [1
]. Models provide data on a given intervention’s impact, cost-effectiveness, and/or financing requirement estimates. Models allow analyses of situations that are difficult or impossible to replicate in real life, including in field trials, such as the absolute impact of a new malaria control intervention in the absence of any existing interventions. They can provide insight by analyzing complex scenarios and identifying which are most likely to occur and which parameters, and their ranges, are the most influential [3
]. It is important that modeling estimates be made available to support evidence-based decision-making. This paper describes a new model for vaccines against malaria, a disease that caused approximately 660,000 deaths in 2010, mostly of children in sub-Saharan Africa [4
]. The Malaria Vaccine Model (MVM) was designed to assist vaccine developers and policymakers in developing countries and in global organizations to make informed decisions about the design and adoption of malaria vaccines.
The demand for evidence on which policymakers and developers can base decisions is increasing. Those developing new, often more expensive, public health interventions for use in developing countries must invest in interventions with the appropriate attributes (e.g., level of efficacy, costs, mode of delivery) to realize desired health impacts. Such decisions will need to be supported by modeled estimates, such as potential impact and financial requirements. The GAVI Alliance (GAVI) has invested close to $100 million since 2000 in activities related to Haemophilus influenzae
type B (Hib), pneumococcal conjugate, and rotavirus vaccines. Establishing the value of these vaccines through the generation of impact estimates was one of the key activities, which arose from the recognition that multi-year delays occurred in the introduction of Hib vaccine by countries, in part because of the lack of data on the burden of disease and potential impact of vaccines [5
]. Ensuring that global, regional, and country decision makers have access to these data in advance could improve the timeliness with which future interventions reach those in need.
A number of models have recently been used to estimate the impact of interventions worldwide. Some were intended to inform global policies and have focused on individual vaccines, such as human papilloma virus, HIV, and rotavirus vaccines [6
]. By contrast, the Lives Saved Tool (LiST) estimated the impact of up to dozens of child survival interventions (including malaria control interventions) across 42 low- and middle-income countries worldwide [9
]. This model, which was intended to help global policymakers prioritize interventions, has been extended for use in individual countries. The ProVac Initiative in the Americas included a model to support country decision-making on the use of new vaccines; the London School of Hygiene and Tropical Medicine and the Swiss Tropical and Public Health Institute (Swiss TPH) developed a web-based tool built on modeled data to assist African policymakers in making local decisions on the use of intermittent preventive treatment of malaria in infants [10
]. These examples illustrate the importance of being clear about a model’s purpose and target audiences, and of taking into account the impact of multiple interventions against diseases.
A malaria vaccine model needs to inform vaccine developers as well as policy and financing decision-makers at global, regional, and country levels. Among other requirements, it needs to be linked to data on transmission and epidemiology in each country and allow for the consideration of malaria vaccines in the context of other malaria interventions available to countries. Malaria vaccine models were recently reviewed by the World Health Organization (WHO) [12
]. There are two published dynamic models that estimate the potential impact of malaria vaccines. One model has been under development since 2003 at the Swiss TPH [13
]. The second, more recent, model was developed at Imperial College London [14
]. Both models consider the dynamics of malaria transmission and of natural immunity to Plasmodium falciparum
using simulation approaches to reflect the underlying relationship between interventions and how much disease they may be able to avert. These approaches can be difficult for non-modeling specialists and policymakers to utilize. The WHO encourages policymakers to include model based cost-effectiveness estimates in support of eventual malaria vaccine adoption decisions [15
]. However, published economic analyses of malaria vaccination based on such models [16
] so far have not considered supply-side considerations, such as manufacturing capacity, which influence implementation and ultimately impact on health.
With global efforts to develop malaria vaccines showing promise, the PATH Malaria Vaccine Initiative (MVI) worked with partners to develop a Malaria Vaccine Model (MVM) that extended the Swiss TPH model by including supply-side considerations and country-specific estimates, and that allowed vaccine developers and policymakers to more easily access the outputs of the Swiss TPH model. The MVM was designed to fill a range of roles for different audiences. For non-profit organizations working on vaccine development, the MVM can help explore the trade-offs in potential product characteristics as part of informing and prioritizing R&D decisions. The MVM can inform global and regional policymakers of the potential impact associated with various delivery strategies and the financing needs for malaria vaccines. At the country level, it can assist local policymakers, for example, through targeted, facilitated workshops, in understanding the potential impact of malaria vaccines in the context of existing malaria control interventions.
The first version of the MVM was developed with the Boston Consulting Group and Swiss TPH and utilized by MVI and partners between 2005 and 2007 to inform discussions regarding the establishment of an Advance Market Commitment for malaria vaccines [17
]. The current version of the MVM, presented here, was developed between 2008 and 2010 by MVI in collaboration with Swiss TPH and Applied Strategies. This version reflects the understanding gained from experience with the initial version about underlying model parameters and design (including outputs), potential uses, and software. The MVM uniquely integrates supply, demand, and cost with public health impact estimates generated from a dynamic malaria vaccine model, while presenting both country-specific and global level data in a user-friendly interface.
This paper describes the major design features, critical parameters, and outputs of the MVM. A demonstration scenario was created that is used to illustrate the model functionality and is not intended to estimate the impact of any specific malaria vaccine. While the modeled vaccine efficacy is substantially higher than the most clinically advanced malaria vaccine candidate today [18
], this demonstration scenario models a hypothetical vaccine based on a strategic goal for the development of a malaria vaccine set by the international community in 2006 through the Malaria Vaccine Technology Roadmap: the development of a vaccine with protective efficacy of more than 80% against clinical disease and lasting longer than four years [19
]. The results of this demonstration scenario are presented, along with lessons learned for the future development of similar models.