UK investments in global infectious disease research 1997-2010: a case study.

BACKGROUND
Infectious diseases account for 15 million deaths per year worldwide, and disproportionately affect young people, elderly people, and the poorest sections of society. We aimed to describe the investments awarded to UK institutions for infectious disease research.


METHODS
We systematically searched databases and websites for information on research studies from funding institutions and created a comprehensive database of infectious disease research projects for the period 1997-2010. We categorised studies and funding by disease, cross-cutting theme, and by a research and development value chain describing the type of science. Regression analyses were reported with Spearman's rank correlation coefficient to establish the relation between research investment, mortality, and disease burden as measured by disability-adjusted life years (DALYs).


FINDINGS
We identified 6170 funded studies, with a total research investment of UK£2·6 billion. Studies with a clear global health component represented 35·6% of all funding (£927 million). By disease, HIV received £461 million (17·7%), malaria £346 million (13·3%), tuberculosis £149 million (5·7%), influenza £80 million (3·1%), and hepatitis C £60 million (2·3%). We compared funding with disease burden (DALYs and mortality) to show low levels of investment relative to burden for gastrointestinal infections (£254 million, 9·7%), some neglected tropical diseases (£184 million, 7·1%), and antimicrobial resistance (£96 million, 3·7%). Virology was the highest funded category (£1 billion, 38·4%). Leading funding sources were the Wellcome Trust (£688 million, 26·4%) and the Medical Research Council (£673 million, 25·8%).


INTERPRETATION
Research funding has to be aligned with prevailing and projected global infectious disease burden. Funding agencies and industry need to openly document their research investments to redress any inequities in resource allocation.


FUNDING
None.


Introduction
Infectious diseases cause a high burden of largely avoidable morbidity and mortality worldwide, and place substantial strain on the limited health budgets, health systems, and economies of aff ected countries. WHO fi gures 1 for low-income countries suggest that infections of the lower respiratory tract are the leading cause of death, followed by ischaemic heart disease, with diarrhoeal disease the third highest and HIV/AIDS the fourth highest cause of death. Although infectious disease control is of the utmost importance for human health, global health security, economic stability, and international develop ment do not have a comprehensive surveillance system to document and monitor infectious disease research investments.
Many factors aff ect the fairly low level of investment in research and development for infectious diseases and maternal, neonatal, and child health. These factors include market failure 2-4 because of low fi nancial opportunities for private investors, risks of research (especially in children and pregnant women), and fragmented infrastructure to do trials for infections and disorders aff ecting populations in low-income countries. Funding for these diseases has been from donor governments, philanthropic organisations, and publicprivate partnerships.
Infections do not recognise borders. Investment in research and development for infectious diseases produces global public benefi ts that have a positive eff ect both locally and worldwide, irrespective of the site of the work or the location of the institution receiving an award, bringing substantial health, social, and economic benefi t. 2,5 In view of the scarcity of resources available, funds for research and development should take into account the local and global burdens of disease. Since 2007, the G-FINDER project, originally commissioned by the Bill & Melinda Gates Foundation, annually surveys global neglected disease research and development expenditure. 6,7 Studies to assess the research spend according to the burden of disease, with data sourced from Australia, Canada, Spain, and the USA, [8][9][10][11] have had diffi culties because of poor data availability. The UK is the second largest investor in global health, but there has been no detailed analysis of its research investment. A study by the UK Clinical Research Collaboration con sidered the broad direction of research funding across all medical specialties, but was restricted to one fi nancial year. 12 Investments from the UK pharmaceutical industry (a key investor in infectious disease research) are poorly documented, partly because of commercial sensitivity. 12 There are large gaps in our quantifi cation of the worldwide spend on infectious disease research and the translation of funds along the research and development value chain into health policy and practice. We present an in-depth analysis examining the investments awarded to UK institutions for all infectious disease research, over the 14 year period from 1997 to 2010. The aims of our study were to quantify awards to UK institutions for local and global infectious disease research; to establish the clinical diseases, specialties, and study types targeted by the major funders; and to identify potential areas of historical and current underinvestment.

Data sources
We obtained data from several sources for studies where funding was awarded between 1997 and 2010. Variables collected included study title, abstract, funding awarded to the study, lead institution, principal investigator, and year of award. 70 principal investigators were contacted individually for further information where needed. We include in the appendix the full list of keywords used to search databases and websites belonging to funding agencies. We also searched ClinicalTrials.gov, the National Research Register, and the Association of Medical Research Charities database.

Inclusion criteria
We used the study title and abstract (where present) as a fi lter, and included all infection-related studies with funding awarded in the period 1997-2010, studies where the lead institution was based in the UK, and infrastructure grants with a clear purpose for infectious disease research. This period was selected on the basis of accurate data availability, which enabled us to suitably compare our results with the WHO burden estimates of 2004 and 2008. We excluded studies not immediately relevant to infection, veterinary infectious disease research studies (unless there was a clear zoonotic component), studies where viral vectors were used to investigate noncommunicable diseases, grants for symposia or meetings, and studies where there were UK contributions (eg, as a collaborator) but the funding was awarded to a non-UK institution. Where there was uncertainty, the study author was contacted or further details were sourced from the internet. Of the studies included in the fi nal database, all had a title and 58% had an abstract. We excluded openaccess data from the pharmaceutical industry since it was clearly under-representative.

Data management
Grants awarded in a currency other than pounds sterling were converted to UK pounds using the mean exchange rate in the year of the award. All grant funding amounts were adjusted for infl ation and reported in 2010 UK pounds. Grants were not modifi ed according to levels of overheads applied to the award. For multicentre studies, any distribution of funding from the lead centre where the award was made to other study sites was not documented. Unfunded studies were excluded from our analysis.
Each study in the database was reviewed by MGH and assigned to as many primary disease categories as appropriate (appendix). Within each category, topicspecifi c subsections were documented.
Studies were also allocated to one of four research and development categories: pre-clinical; phase 1, 2, or 3; product development; and operational research (appendix). All studies were categorised and subsequently double-checked by JRF. Provisional datasets were circulated to all authors for review and comment. MKC and FBW further verifi ed a random sample of 10% of the data (663 studies) in a third round of checks. The fi xed marginal κ score was 0·950 suggesting high agreement between the authors when categorising studies. All diff erences on inclusion or categorisation were discussed between MGH and the author who fl agged the study, and where there was still disagreement, the data were     and specifi c keyword queries used to select precise sections of the data for analysis. Statistical analysis and generation of fi gures and graphs were done with Stata (version 11). Regression analyses were reported with Spearman's rank correlation coeffi cient (r) to establish the relation between research investment, mortality, and disease burden as measured by disability-adjusted life years (DALYs). 13

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 fi nal responsibility for the decision to submit for publication.

Results
We identifi ed 320 470 studies that were suitable for screening. Of these, 6170 funded studies met our inclusion criteria (fi gure 1). The funding for these studies represented a total research investment of UK£2·6 billion (fi gure 1). Worldwide, gastrointestinal disease represents a similar burden of mortality and DALYs as HIV but receives roughly half the research funding. 1 Investment is similar when classifi ed by research themes, health burden, and research. Table 1 shows a detailed breakdown of the investment by disease and cross-cutting themes, and the burden data. There are prominent disparities between investment and burden of specifi c diseases. The type of science funded by each institution   Regression analysis between disease system and infection against DALYs shows a clear misalignment between investment and worldwide burden (fi gure 2), with a moderate association of specifi c infection research funding to DALYs in 2004 (r 0·5270) and a worsening association in 2008 (r 0·3203). Conversely, there is a positive association between infection by disease system and DALYs in 2004 (r 0·8810) and 2008 (r 0·8333). Considering their burden according to DALYs in 2004 and 2008, trachoma, syphilis, and gonorrhoea are among the infections that are most underfunded, relative to all infections in the study. Hepatitis C, African trypanosomiasis, leishmaniasis, and malaria are among the infections that are most overfunded.
When analysed by research and development pipeline (fi gure 3), £1·6 billion (62·5%) of the investment was allocated to preclinical research, with smaller amounts (£146·8 million, 5·6%) allocated to clinical trial research  Sources of funding by research phase vary greatly (table 2). The Medical Research Council was the leading funder of HIV and virology research, with £360 million of investments (36·0%). The Wellcome Trust is the leading funder of malaria and parasitology with £275 million (41·7%), as well as bacteriology research, with £176 million (30·1%). The BBSRC is the leading funder of mycology research with £14·8 million (30·5%) and the UK Department of Health is the leading funder of research into prion disease with £20·2 million (60·2%).

Discussion
We identifi ed 6170 funded studies, with a total research investment of £2·6 billion. Studies with a clear global health component represented 35·6% of all funding (£927 million), and the Wellcome Trust and Medical Research Council were the two leading funding sources. Preclinical research accounted for £1·6 billion (62·5%) of the total research and development invest ment. We highlight several major areas where there might be underinvestment-namely, for research focusing explicitly on infections in elderly people (£7·2 million, 0·3%) and in children (£87·1 million, 3·3%). Investment in some of the neglected tropical diseases, gastrointestinal infections, and sexually transmitted infections excluding HIV are also far lower than their global burden of disease would warrant.
Investment for drug-resistance-related research seems inadequate, since antimicrobial resistance across all areas of infection has been described by WHO as a global public health emergency aff ecting all countries. 14 Despite the expansion of the directly observed treat ment shortcourse (DOTS) programme for tuberculosis, multidrugresistant tuberculosis is spreading unabated such that WHO now considers strains of extensively drug-resistant tuberculosis to be "virtually untreatable". 15 Future research investment in this area should increase in line with projected burden. WHO data suggest that gastrointestinal infections and diarrhoeal disease account for high disease burden and mortality worldwide. 1 However, the research spend for these disorders is substantially lower than other high burden and high profi le diseases, such as HIV and malaria. In view of their relative burden, gastrointestinal infections,  and sexually transmitted infections should therefore be assigned a higher priority for improved research investment by funding organisations. There might be a reasonable argument for a proportionate increase in research related to health-care-associated infections, although there is also a lack of good quality data about burden in this area, particularly in low-income countries. The burden of infectious diseases is particularly heavy on children, with 64% of worldwide deaths in children younger than 5 years related to infection. 16 We show two key shortcomings for global infectious disease research in relation to children. First, investment in research for infections in children is generally very low, although malaria and vaccines, which mostly relate to children, attracted fairly high levels of funding. From 1997 to 2010, studies specifi cally focusing on infectious diseases in children attracted £87·1 million across 307 studies, representing 3·3% of the total funding across 5·0% of the total studies. Second, studies relating to nutrition and paediatric infectious diseases were poorly funded despite the huge burden of morbidity and mortality worldwide in children due to malnutrition-a major cause of immune defi ciency. 17 Paediatric infectious disease studies with a nutrition component attracted £4·3 million across nine studies, representing only 5·0% of paediatrics funding and 0·2% of total funding. Analysis by research and development value chain shows that the UK has invested heavily in preclinical research, but invested relatively small amounts in phase 1-4 trials or product development. This discrepancy might represent a strength of UK institutions in preclinical science, but it also highlights a need to strengthen research capacity further along the research and development value chain. There is also a need to obtain comparable data from other countries to under stand whether this spending pattern is representative. It would also be useful to gauge whether funders consider they receive a lack of high quality clinical grant appli cations compared with those in basic science. We noted a lack of readily available data from the pharmaceutical industry, greatly underestimating funding for clinical trials of pharmaceutical products; but this will probably make little diff erence to estimates of funding for operational research.
Linking investments to disease burden to optimise the allocation of limited research funds is a challenging endeavour. Our fi ndings could be used to develop transparent and objective methods to better couple research investment to burden of disease. Earlier analyses have broadly concluded that the fi nancial spend is appropriate when DALYs are used as a measure of burden, but using measures of incidence or prevalence as a marker for burden of disease were insuffi cient [8][9][10][11] in view of the unreliability of incidence data and that improved management of infections with a high mortality rate often lead to an increase in prevalence. 18 However, defi ning an appropriate amount of research investment for each disease category is challenging, since diff erent levels of investment might be needed to address diseases with a similar level of public health burden. Emerging infections with unpredictable future disease burden such as prion disease, viral haemorrhagic fevers, or pandemic infl uenza present particular challenges when establishing relative priority for investment.
Our fi ndings are consistent with earlier studies that focused mainly on research and development spending for global infectious disease, 7 showing the UK to be a leading funder of research and development, along with the USA and European Commission, and showing private sector contributions to neglected disease research to be an estimated US$503 million (£325·4 million) in 2010. Although there is no breakdown by country, the data are categorised by disease area, with tuberculosis, malaria, and dengue attracting the most private investment. Investment by the pharmaceutical industry could aff ect how other funders invest in research and development for infectious disease (no industry data is included in this analysis because of diffi culties in openly accessing funding information), 8 whereas research charities have their own specifi c areas of focus that might constrain their ability to invest beyond selected diseases. A study of offi cial development assistance allocated for neglected tropical diseases shows low investment levels accounting for only 0 . 6% of annual health assistance between 2003 and 2007. 19 Tracking the overall spend on all areas of global health fi nancing is a complex task. This conclusion was based on several factors including fragmentation of data and paucity of detailed information from the private sector. A Global Health Resource Tracking Working Group reported in 2006 that calculating the amount of funding allocated to global health was too diffi cult owing to several factors including tracking the large and diverse number of public and private sources of funding, and the nature of poorly designed donor accounting struc tures. 20 A 2009 study investigating global health funding recommended the provision of detailed descriptions of the funding provided to improve the effi ciency, accountability, performance, and equity of resource allocation of the many actors that populate the global health landscape. 21 One key recommendation of the Global Health Resource Tracking Working Group was to implement improved tracking and monitoring of global health fi nancing. Within research, this improve ment can be achieved both worldwide and nationally, as earlier studies and our report show in the area of infectious diseases.
An important question inspiring our project is whether the right research is being funded. Although the competitive research process used by most funders when awarding grants can help ensure that the funded portfolio is of a high quality, absence of explicit resource allocation criteria means funding for research and development might not reach areas of highest burden. Funding agencies do have their own areas of focus, and thus UK funding agents might have factored other coun tries' investments into their own investment strategy. Data from other countries is essential to complete the mapping of investments. WHO budgets and global disease burden have been the centre of much debate. [22][23][24] The low profi le of neglected tropical diseases despite a high disease burden against other tropical diseases such as malaria has been highlighted. [25][26][27] Studies have also explored ways to maximise the eff ect of operational research on policy and practice. 28 Our fi ndings build on these earlier studies and contribute to policy discussions relating to investment in research. They also inform funders of funding patterns among organisations fi nancing re search and development, which can help prevent sub optimum investment of limited resources.
Showing the relation between health burden and research funding allows identifi cation of areas of underinvestment. However, we cannot state with certainty that these gaps represent areas of neglect without factoring in several considerations. These factors include the feasibility of doing the research, the cost of the technologies involved, the presence or absence of suitable infrastructure and appropriately skilled individuals, local political and social conditions, and uncertainty around the accuracy of the estimates of disease burden.
Our analysis has several limitations. We rely on the accuracy of the original data from the funding organisations; although checks were made on any apparent discrepancies or obvious errors, any interpretation of these original data is potentially fl awed. No attempt was made to investigate any contribution of indirect and estate costs (including the introduction of full economic costing formulae in the UK), and currency conversions for donations in US dollars or Euros might not be precisely representative of the funding awarded because of fl uctuations in fi nancial markets across 1 year. Unless the information was clearly documented, we do not have data to assess how much funding was distributed from the lead institution to study partners. Study numbers will be slightly inaccurate owing to diffi culties ascertaining whether the funding was related to project extension or new study, and whether the funding was for a site as part of a multicentre study. Diff erences in study reference numbers were used as a guide to distinguish between new studies and extensions and eff ort made to identify multicentre trials.
Details of private sector research funding are diffi cult to obtain and analyse in the level of detail we were able to apply to data obtained from public sector and charitable foundations. The National Research Register lists awards of research of direct relevance to the NHS from 1997 to 2007. The register closed at the end of 2007. We could not openly access data for pharmaceutical industry in-house research and development investment, since much of this information is considered commercially sensitive. Individual awards of many millions of pounds for research into specifi c diseases could skew the results. There are no data from the Chief Scientist Offi ce (Scotland) from 2008 to 2010, which might underestimate overall fi gures for research and development.
We cannot ascertain what proportion of a grant should be allocated to each of the allocated disease categories. Hence, there might be disagreement about how the categories have been assigned. Some studies could not be allocated to categories since there was no clear implication of association with that category-for example hepatitis B could not be allocated to sexually transmitted infections unless this factor was suggested in the study title or abstract, owing to the pathogen's many modes of transmission. Creation of disease categories and allocation of studies to the categories is subjective.
Burden measures are typically an estimate, since data could be missing, unobtainable, or subject to a diff erent classifi cation system or case diagnosis. Our analysis cannot easily account for the cross-disciplinary or geographical eff ect of research.
Our report presents the latest fi gures on investments in local and global infectious disease research awarded to UK institutions between 1997 and 2010 (panel). Neglected tropical diseases, gastrointestinal infections, sexually transmitted infections, and antimicrobial resistance seem to be areas warranting increased investment. We will make the entire database and fi gures available online to assist policy makers, funding organisations, and fellow researchers in the identifi cation of research gaps and infectious disease priorities (see margin link). We urge funding organ isations to make their investment portfolios

Systematic review
We searched PubMed for articles published at any time with the search terms "investments in research" and "infectious disease burden", as well as "burden" and specifi c infectious diseases or cross-cutting themes. We also searched for published reports with the same search terms. No publication investigating the investments in infectious disease research by the UK over time was identifi ed.

Interpretation
Our study is the fi rst to do a detailed assessment that the infectious disease research investments made by funding organisations to UK institutions. Health research investment decisions need to balance strategic insight into the burden of disease with judgments on scientifi c quality and novelty. We identify inconsistencies of investment compared with the global burden of infection, suggesting the need for strategies to redress this imbalance. The scientifi c and public health community, as well as governments and health departments, need to ensure limited resources are allocated appropriately and strategically. We encourage the support of similar open-access databases for non-communicable diseases and other countries, as well as further work comparing research funding with disease burden.
For the Research Investements in Global Health database see http://www. researchinvestments.org openly accessible on this website by reporting their successful grants each year, as we have seen with the clinical trials registry ClinicalTrials.gov. We encourage the develop ment of similar databases for non-communicable dis eases and other countries, as well as further work on comparing research funding to disease burden.
High-quality research can allow substantial improvements in redressing the infectious disease burden. As emphasised by the neglected tropical diseases movement, increased funding and better-informed resource allocation could help control, eliminate, and eradicate infectious diseases. 29 The scientifi c and public health community, as well as governments and health departments, need to ensure limited resources are allocated appropriately and strategically, particularly with regards to health care and the alleviation of infectious disease morbidity and mortality.