Data sources and collection

We sourced information from websites, funding organization databases, and the published literature. Data from the UK, European Union (EU) and the US were included for the period of 1997 to 16 April 2015, which represented a 12–month period following the announcement of the outbreak by the WHO. Variables collected included study title, abstract, website, grant type, funding awarded, name and gender of the principal investigator, host institution, year of award and projected duration of project.

In the UK, research funding organisations included the UK Medical Research Council (MRC), the Wellcome Trust, the UK Department for International Development (DFID); in the EU organisations included the European Commission, the European Centre for Disease Prevention and Control (ECDC); and in the US organisations included the National Institutes of Health (NIH), Congressionally Directed Medical Research Programmes (CDMRP), Bill and Melinda Gates Foundation, Paul G. Allen Foundation, Burroughs Wellcome Fund, Doris Duke Charitable Foundation, Howard Hughes Medical Institute, Donaghue Foundation, Ellison Medical Foundation, Arcus Foundation, and the Roy Carver Charitable Trust. The organisations selected represented the leading infectious disease and public health research funders in the respective regions. Variables collected included financial disbursements, project title, website, donor organization, recipient organization, recipient country, and year of award.

Data management

All grant funding amounts were reported in 2015 US dollars (US$). All awards were adjusted for inflation and converted to US dollars, using the mean exchange rate in the year award (http://www.oanda.com). Grants were not modified according to levels of overheads applied to the award. For multi–center studies, the distribution of funding was accounted for, where openly available. Unfunded studies, or studies without a clear funding amount, were excluded. Private sector sources of funding were excluded, as data was not openly available. Channels without a robust data source were excluded from the final analysis.

Data were collected over a period of 7 months, from October 2014 to April 2015. The study title, abstract, and website were used to filter and categorise research studies. All research studies were reviewed by two or more co–authors. Each research award was allocated to one of five R&D categories along with the research pipeline: preclinical research, phase 1, 2, or 3 clinical trials, product development, public health research, and cross–disciplinary research. Public health research included surveillance, epidemiology, modeling, bioinformatics, and operational research. Cross–disciplinary studies were large–scale projects, with significant funds to facilitate two or more subprojects to work in parallel. Duration of research studies was also estimated. Top 3 donors were ranked for each recipient organization.

Statistical analysis

Microsoft Excel 2011 (Microsoft, Seattle WA, USA) was used to categorise the research studies and generation of tables. Statistical analysis and generation of figures and graphs were generated using Stata (version 11.2) (STATA Corp LLP, College Station, Texas, USA). Simple regression analyses were reported using Spearman’s rank correlation coefficient (ρ), to assess the degree of correlation between research investment, disease burden, and bibliometric impact. We used fold differences to compare total investment, number of studies, mean and median grant size.

Role of the funding source

There was no funding source for this study. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Implications of mapping research funding

Emerging infections have the potential to disrupt the global economy and global health, mobilising significant resources over short periods of time. In an effort to learn from the latest Ebola outbreak in West Africa and prevent the next epidemic, data on current investments, coupled with data on disease burden and efficacy of interventions, may provide an enabling environment to strategically allocate scarce resources [15].

We demonstrate significant public and philanthropic funds have been invested in Ebola and Marburg virus research in 2014 and 2015.

Evaluating the impact of R&D is not only good science, but also a social imperative. The responsibility to invest appropriately lies both with researchers, and with funding organisations. Policymakers must work closely with the scientific and funding communities to facilitate channels that provide both the flexibility and the strategic resources to alleviate the burden of emerging infections. Evidence–informed investment is key to allocating resources wisely and fairly. Our analyses provide a first step in aggregating and describing the trends with Ebola and Marburg viruses, in an effort to develop more sensitive methods to evaluate the public health impact of supported research.

What the Ebola crisis in West Africa has demonstrated is the immense repercussions of an outbreak on the stability and social cohesion of a society. The spread of the outbreak is linked to the under–developed and under–resourced health systems preexisting within these countries, but also the imbalance between investments in the diseases that affect individuals in wealthier countries over those living in low–income settings [16].

Proposals for research strategy

New approaches for research during a new epidemic or pandemic are critical, and useful examples include the 2009 NIHR “Rapid Response” research funding for influenza [17], and the 2011 NIHR “Sleeping grants” where protected funding will be activated in the event of a new pandemic [18].

A portfolio of research along the R&D value chain (extending from preclinical and laboratory science, phase 1 and phase 2 clinical trials, large–scale phase 3 and phase for clinical trials, translational studies and evaluations) is required to address the greatest challenges in global health. Investing in the progression and links between these different types of research is essential in order to build on early stage research findings, and translate breakthroughs in the laboratory into reality in health systems where translation and uptake of innovations remains a challenge [19,20]. Operations research is critical to understand how innovations can be effectively scaled up [21,22].

Encouragingly, the scope and volume of potential Ebola diagnostics, therapies and vaccines is broadening ( Box 1 ).

R&D should also extend beyond vaccines, diagnostics, and antivirals, however. Research into novel, digital surveillance systems is also warranted. Surveillance systems are essential to detect outbreaks of emerging infections, and to mitigate their health, security and economic effects, in all countries [27,28]. The International Health Regulations (IHR) confers responsibilities to WHO member states to develop surveillance systems to detect and respond to public health emergencies, although many are not ready [29]. IHR provides a mandate for global surveillance, a safety net to detect disease outbreaks if they are not detected and/or reported by countries. Responsive health systems and global collaboration is critically important if these infections are to be rapidly contained [30,31]. In 2002, the Global Public Health Intelligence Network, developed by the Public Health Agency of Canada, openly alerted the global health community to SARS 2 months prior to the WHO. In 2014, HealthMap sounded the alarm to an outbreak of unknown etiology in Guinea 9 days before the WHO. Recent work has further highlighted the density of Google Trends searches for Ebola in West Africa, despite inadequate Internet coverage in the region [32]. These factors together highlight the need for research into digital systems for global health security.

Limitations and scope

The primary limitation to this work is the difference in quality of research funding, compared with emergency response funding. Although public and philanthropic research funding is, on the whole, well documented and often openly accessible online, there are clear quality concerns with emergency response funding. For instance, it is difficult to verify disbursements and pledges for emergency response, and this is a notable problem with the latest Ebola outbreak. As published in the Lancet Global Health – the tools “for tracking resources in a crisis are not fit for purpose” [33]. Despite the challenges with emergency response funding, it is important to document the openly published information, systematically analyze the data and start a discussion on next steps in tracking resources during a crisis. We need a robust financial platform to monitor and evaluate investments – with the capacity to cross–talk with research funding commitments. With the re–emergence of Zika virus in the Americas, this needs to occur without delay.

There are also limitations to analyzing global expenditure data search, elaborated in greater detail in a recent article by Young and colleagues [34]. The first is the challenge of comparing data collected from different countries. The Frascati manual attempts to provide some guidance on the data collection and currency adjustments for R&D [35]. Despite some clear guidance, two recent studies highlight the ongoing difficulties with methodology and subsequent impact on interpretation, particularly due to discrepancies in data conversion [36–38]. The second is the lack of indices to convert health R&D expenditure to a single, commonly used currency [38]. One of the main reasons this is important is the comparative difference in the costs of input for R&D in different settings [39]. There are two possible ways of converting to a common currency: current exchange rates, and GDP purchasing power parities (PPPs) [40]. The Frascati manual recommends GDP PPPs, and current exchange rates may underestimate the contributions of countries such India and China, by overstating the costs of R&D. The third is the need to adjust investment over time, in order to account for the potential role of inflation. There is currently no index specific for health R&D and the Frascati manual advocates for using the GDP price index, with the limitation that is not specific for R&D. The special R&D deflator is an index adopted by the US National Institutes of Health and may be adopted for high–income settings [41]. Using this index for settings with high inflation will overestimate respective contributions, and vice versa. The fourth challenge relates to the need first to deflate data in national currency, then to covert the adjusted data to a common currency using a selected base year [42]. Altering data using another mechanism may skew the data inappropriately and alter the comparability between settings. Fifth, our study is likely to underestimate research funding, partly due to the lack of openly available data from the pharmaceutical industry. Sixth challenge relates to difficulties in ascertaining the right proportion of a grant allocated to a specific disease, when there are multiple research sites and conditions researched, which is not uncommon for co–infection studies.

Strategic coordination of research funding

A Global Observatory on health R&D has been proposed by member states of the sixty–sixth World Health Assembly in 2013 to “provide a mechanism to monitor and analyse relevant existing information on health R&D, including resource flows, product pipelines, and research outputs, with a view to contributing to the identification of gaps and opportunities for health R&D and to inform priority–setting for new R&D investments based on the public health needs of the world’s poorest countries” [43,44].

A clear, open source repository for R&D investment data could address the problem by showing the funding landscape of both clinical and non–clinical studies in real time [45]. Gaps in the data highlight the need for a global health R&D observatory. Innovations in reporting will help improve priority setting to address burden of disease in low– and middle–income settings. Interpreting disease burden data in the context of R&D funding data are an essential step in the equitable allocation of health investments. Whether the observatory should be run by the WHO or by another institution or consortium, however, remains for debate.