Is research in health aimed at the most important problems? Are resources assigned efficiently? Do spheres of study obtain more or fewer funds for healthcare reasons or is it a question of “fashions”? Comparisons between the reach of research (resources, budget, etc.) and the personal, economic and social costs of each disease, point to the existence of substantial misalignments, at both global and local levels. According to the importance of their effects, some diseases that are prevalent in low- and medium—income countries, such as malaria and tuberculosis, may deserve more investment in R&D. This is also true of diseases such as depression and stroke that, although affecting hundreds of millions of people all over the world, attract relatively little research. This type of comparative approach to health needs and R&D efforts can be useful when prioritising lines of research. Consulting healthcare data more profusely and taking patients and citizens more into account should both be indispensable strategies for assigning health research resources efficiently.

People’s health was vastly improved over the course of the last century, largely due to improvements in living conditions, sanitation (e.g. access to clean water) and healthcare, but also, to a lesser extent, to scientific discoveries. X-rays and penicillin are examples of  breakthroughs that led to great medical advances, while vaccines are the fruits of applied research that has saved millions of lives.

Such extraordinary successes have led to the belief that more biomedical research will inevitably lead to better health. However, the reality is not so simple. If biomedical research is to improve health it must first take into account factors that, strictly speaking, lie outside the realm of medicine. First and foremost more attention must be paid to people’s needs and also to the causes of ill health (for example, broader environmental and socio-economic conditions such as pollution and inequality). Taking this into account, it is then possible to establish more effective R&D priorities.

In 2010 the world spent an estimated 240 billion US$ (purchasing power parity) on health R&D; 90% of this sum corresponded to high-income countries. Health R&D expenditure has grown over recent years, coming to represent between 0.2% (0.26% in Spain) and 1.0% of GDP in high-income countries (Graph 1: Røttingen et al., 2013). 

Sixty percent of these R&D investments were made by the private sector, 30% by the public sector and 10% by other sources, such as non-profit organisations. While acknowledging the past and potential future benefits of health research, in recent years various analysts have questioned the criteria being used to decide priorities for research and the assigning of resources.

Specifically, most attention has been aroused by the problem of “translation”, i.e. whether the knowledge derived from fundamental science research is quickly and efficiently  applied in clinical practice and healthcare. Accordingly, many ‘translational’ centres and labs have been created with the aim of combining basic and clinical research, and thus quickly transferring new knowledge ‘from bench to bedside’.

If we take a step back, however, and look from a wider perspective, we find deeper problems with current biomedical research: An alarming proportion of scientific discoveries published have proved false (Ioannidis, 2005), while a significant number of experiments cannot be reproduced (i.e. the same results are not obtained when repeated) and there are also instances of publications unnecessarily replicating previously known findings.

Current evaluation systems are mainly focused on scientific visibility (e.g. publication in ‘top’ scientific journals), which may be leading research away from pursuing knowledge that is valuable to society. Scientists are shifting their research areas and theories to the issues that are published in these journals, even though they might not be the most relevant socially. And perhaps the most serious and widely discussed issue is that private R&D (60% of total health R&D) follows the demands of the market – which focuses on diseases in rich countries, particularly chronic conditions - while diseases or findings of greater public health relevance continue to receive insufficient investment.

In the light of these dilemmas, many analysts argue that health benefits from research could be significantly improved by more systematic planning of research priorities. Research funding is allocated for various reasons, generally based on criteria that include perceptions of the scientific quality of the projects and teams involved, the potential for scientific advances and society's demands or needs for a given issue. In practice, however, although great emphasis has been placed on assessing scientific quality so as to foster excellence, far less attention has been paid to assessing whether research addresses social needs. Recent studies suggest there is a lack of alignment between research priorities and health needs, both at national level and, even more so, in global health research (Gross et al., 1999; Røttingen et al., 2013; Evans et al., 2014).

In this article, we will analyse how decisions on establishing priorities in health research can be improved by asking whether its contents respond or not, and to what extent, to social needs. To do this, we compare evaluations of health needs with proxies of research initiatives. The significant misalignments observed support the view that health research can greatly benefit from two things: extensive and systematic use of all the data stored in the healthcare system (through processes that exploit ‘big data’), and greater participation of the diverse social agents. This shift would align with ongoing policy discussions of ‘open science’ and Responsible Research and Innovation.

The ultimate goal of health research is to improve the health and wellbeing of humankind and it should, therefore, be focused on this goal. When setting public policies and research priorities it is usually stated that health research should be orientated ‘towards humanitarian goals and solidarity’, which should ‘underpin every step in the innovation cycle from discovery in bench and bedside research, to implementation in healthcare, and prevention.’ (EC, 2016). But how can we know if such goals are indeed fulfilled in the long term?

To try to answer this question, in 1997 the US National Institutes of Health (NIH) created a working group on the establishment of priorities for research in the health sphere. This panel of experts concluded that society’s health needs were one of the criteria used to decide which research would be conducted. However, it also warned that the NIH did not sufficiently explain how health needs were assessed and proposed improvements for its analysis of health data. Two years later Gross et al. (1999) published a study comparing various disease burden indicators to NIH research initiatives for a list of major health conditions. Research initiatives were classified based on the funding allocated to each project and disease.

The ‘disease burden’ gives a clue regarding a health condition’s social importance. The value of this indicator is established taking into account a set of parameters; among them, prevalence (the number of people with a given disease), incidence (frequency of new cases), hospital days, mortality, years of life lost and disability-adjusted life years (DALYs). DALYs is a measure that estimates how many years of healthy life are lost to disease, taking into consideration mortality as well as disability caused by a health condition. In other words, DALY expresses the number of years of life lost due to a disease, taking into account both life expectancy and the years not fully enjoyed precisely due to that condition (years lost due to disability).

The study by Gross and colleagues (1999) found that there was a weak correlation between funding for the study of a health condition, on the one hand, and mortality and years of life lost on the other. However, they also observed a strong correlation between funding and DALYs. This meant, in short, that NIH funding was, effectively, sensitive to disease burden. In other words: decisions on what to research and how much effort to devote to it took into account the importance, in social and personal terms, of the diseases.

However, this careful analysis also showed substantial disparities in the amount of funding for the same burden across diseases. Diseases such as AIDS, breast cancer or diabetes received, proportionally, much more funding per DALY than depression, colon cancer or perinatal conditions. Consequently, there was room for improvement in the alignment between funding and disease burden. Subsequently, various studies at a national scale have followed this methodology, for example in the US, Norway and Spain. 

Another approach to assess research funding against disease burden has been to use the number of publications as a proxy of research initiatives (Evans et al., 2014). This bibliometric approach has the advantage of embracing research conducted in universities and hospitals by lecturers and physicians that is not covered by the R&D expenditures of funding agencies. In this article, we use indicators on publications to link the disease burden (DALYs) to research initiatives. Data from the PubMed and Web of Science databases corresponding to the period 2009-2013 are combined. The relationship between the publications and diseases is based on PubMed descriptors provided by experts at the US NIH National Library of Medicine. The global burden of disease estimates come, lastly, from the WHO[1] , using 2012 data.

This methodology presents significant limitations and the results have to be interpreted with caution. First, publications do not accurately reflect research initiatives. They fail to detect most private R&D initiatives (which are confidential until their results are patented). Also, due to different disciplinary traditions and incentives, not all research fields publish with the same frequency and certain topics become temporarily fashionable or unfashionable for reasons unrelated to both science and health.

Secondly, there is a very long time lapse between research emerging and its publication (between 2 and 5 years), and an even longer one (5-20 years) between the publication of conclusions and their application in social contexts. There is, therefore, some uncertainty about the appropriate time window for the purposes of comparison between disease burden and research effort. Additionally, since the disease burden also changes over time, future studies should conduct a dynamic analysis of the evolution of DALYs and publications. This way, it would be possible to establish with greater precision the convergence (or divergence) over time between one and another.

If, for one set of diseases, the relative health impact of one of them, measured as a percentage of DALYs, is compared with the volume of research publication generated by it, also measured as a percentage of the total publications for the set of diseases, then it is possible to get an idea of the degree of alignment or misalignment between the disease burden and the research effort (graph 2).

We present data for the whole world and for Spain (which has a profile similar to that of many developed countries). Disease burden and publication patterns show striking disparities in research across various disease groups. At the global level, cancer (malignant tumours) accounts for more than 22% of global disease publications, even though its disease burden is less than 10% of the total. Cardiovascular, infectious and parasitic diseases represent more than 16% of the burden but less than 10% of publications. Neonatal conditions are an issue with very little research in relation to their burden. 

In Spain and other developed countries, disease burdens are strikingly different to those of the world as a whole. For example, cancer’s share of the burden in Spain is a little higher than its share of publications. In contrast, the share of publications on infectious and parasitic diseases is much higher than would correspond to the disease burden of these conditions. 

Graph 3 focuses on specific conditions rather than disease groups, i.e., rather than viewed as a single disease, cancer is broken down into types. The horizontal axis represents the disease’s relative burden (percentage with respect to total) and the vertical axis the proportion of publications (also relative to the total number of publications on diseases). For diseases that appear above the 45-degree sloping line, there are relatively more publications than their disease burden; in other words, they could be classified as “over-studied”. Those that appear under that same line have, in contrast, a proportionally lower level of publications; therefore they could be classed as “under-studied.” 

The comparison between the patterns for the world and for Spain reveals that some diseases have similar burdens in the developed countries (e.g. cardiovascular or cancers), whereas others have a much higher burden in low- and middle-income countries than in high-income countries. Some of these last diseases are still present in high-income countries (e.g. AIDS/HIV or tuberculosis), whereas some other diseases (such as malaria) are only a major burden in poorer countries.

Given these differences, diseases have been classified into three types: type I (disease burden is no more than 3 times higher in low- and middle-income countries than in high-income countries), type II (between 3 and 35-times higher in low- and middle-income countries) and type III (more than 35 times higher in low- and middle-income countries) (Røttingen et al., 2013). Type II and III diseases are shown as red triangles in graph 3.

The greater part of the world’s health research is published in high-income countries, mainly Europe, the USA and East Asia (Graph 4). On the contrary, Africa, the Indian subcontinent (South Asia) and South East Asia have a small percentage of publications in relation to their populations. In high-income countries, there is an imperfect but significant correlation between national disease burden and research initiatives (as can be seen in the case of Spain in graph 3). Since 90% of the world’s research is produced in developed countries, it follows that global research initiatives are more aligned with the health needs of affluent societies than with those of the world as a whole (Evans et al., 2014). One can appreciate this pattern by looking at type II and III diseases (Graph 3, shown as red triangles): in the world, they appear as diseases with high disease burden and little research, whereas in Spain they appear mainly as diseases with little burden but generally more research than the country needs.

In an increasingly interconnected world, there are good reasons for high-income countries to contribute to global health research. In the first place, equity in health is widely acknowledged as a shared value across the world – and governments can contribute to it by supporting research into global health. Secondly, as a result of globalisation and increased movement of people across continents, the ‘globalisation’ of lifestyles and infectious diseases has effects on health across the globe. Under these circumstances, it is in everybody’s interest to improve health conditions on a global scale. Moreover, we have to take into account that private health research (which represents about 60% of global health funding) is even more narrowly focused on chronic diseases of the rich.

To address, even partly, the unmet health needs of low-income countries, governments are now being recommended to spend (depending on sources) between 0.05% and 0.20% of their GDP on health R&D, and at least 0.01% on research (Røttingen et al., 2013). At present, only a few countries (graph 1) meet the first criterion of total health R&D, but in 2010 none met the criterion of 0.01%. (To cite some of the highest spenders: the US spent 0.0096%; the UK, 0.0073% and Sweden 0.0041%. Spain, for its part, spent 0.0010%).

Thus, there are good reasons for transforming public health research for the purpose of it supporting unmet health needs both locally and globally. This calls into question both the reasons for the relative under-investment, in local and global contexts, in diseases such as depression or chronic obstructive pulmonary disease, and the reasons for the excessive effort invested in researching local health priorities (such as diabetes) rather than global health challenges (such as malaria).

Furthermore, the establishment of R&D priorities also has to take into account broader contextual considerations, such as the availability of proven therapies and expert opinions on potential advances that may realistically be reached in fighting a disease. It could be argued, for example, that drugs are already available to lower the high disease burden of infectious diseases in Africa – and hence the problem lies not in lack of research but in inefficient healthcare services and unavailability of affordable essential drugs.

At the other extreme, some experts may argue that, for conditions such as Alzheimer's disease, research should be restricted because the field is not yet mature enough to yield medical solutions. In reality, it has been noted that some diseases are more difficult to address than others via clinical approaches (such as vaccines) and that other types of sociotechnical interventions can be more successful than conventional therapeutic approaches. For example, insecticide-treated bed nets have proven very useful in the fight against malaria.

In summary, the question is not only whether public R&D is tackling the most pressing health needs, but also whether the strategies pursued are appropriate for addressing these needs. Since research is a highly uncertain process, and various R&D strategies may contribute in different ways to improving a health condition, it is advisable to pursue several research lines in parallel (graph 5).

The size of the nodes in graph 5 shows the proportion of publications on a particular topic. This proportion is a proxy of the amount of research resources spent. Is this a wise distribution of resources on obesity? Should scientific policy foster more research into certain topics (e.g. diet) than others (e.g. bariatric surgery), given their relative public health benefits? In the case of obesity, many stakeholders agree that more research is needed into the social and psychological factors related to the food industry, consumption patterns and sedentary lifestyles (graph 5), rather than furthering our understanding of biology or improving therapeutic approaches such as surgery.

In a case such as obesity, it is important to take into account the valuable experiences and informed preferences of patients and other relevant citizens affected by the problem when assessing the most relevant approaches. Given that many prevalent health conditions (such as depression or respiratory problems) are associated with environmental and social factors (such as pollution, urban lifestyles and inequality), engagement is particularly important to appreciate citizens' assessment of health conditions and determinants. Data on healthcare, including global disease burdens, offer new opportunities to spot problems that deserve attention, but citizens' engagement is also necessary to highlight their preferred types of solutions.

This article shows how a comparative approach can help foster a better alignment between R&D initiatives and health needs. The data suggest that some diseases prevalent in low- and middle-income countries, such as malaria or tuberculosis, deserve more R&D investment – but this is also the case for conditions such as depression or stroke that have a high disease burden across the world. While there can be good scientific reasons for some areas having relative over- or under-investment in relation to their disease burden, the comparison proposed here could encourage a better distribution of research resources.

The analysis can be enriched by considering in greater depth the design of more efficient distribution formulas for investments across research portfolios – i.e. by considering the various types of scientific approaches to tackling a disease. Given the enormous uncertainty surrounding research outcomes, the study of a given disease should include a variety of strategies, from basic biology to medicine, and from social studies to public health considerations.

In order for health R&D to be more sensitive to social needs, the process of establishing priorities should be enhanced by incorporating information about health needs and by dialogue with patients and various stakeholders. In accordance with Open Science and Responsible Research and Innovation, the establishment of priorities should be based on two main strategies: first, the collection of the data available on the research initiatives undertaken to address a specific health problem; second, an appraisal of the preferences of citizens on the problems that need to be addressed, and the best strategies for tackling them.

Cassi, L., A. Lahatte, I. Ràfols, P. Sautier and É. de Turckheim (2017): «Improving fitness: mapping research priorities against societal needs on obesity», Journal of Informetrics 11(4).

European Commission’s Scientific Panel for Health (SPH) (2016): «Better research for better health. A vision for health and biomedical research from the Scientific Panel for Health»

Evans, J.A., J.M. Shim and J.P.A. Ioannidis (2014): «Attention to local health burden and the global disparity of health research», PLOS One, 9(4).

Gross, C.P., G.F. Anderson and N.R. Powe (1999): «The relation between funding by the National Institutes of Health and the burden of disease», New England Journal of Medicine, 340(24).

Ioannidis, J. P. (2005). Why most published research findings are false. PLoS medicine, 2(8), e124.

Røttingen, J.A., S. Regmi, M. Eide, A.J. Young, R.F. Viergever, C. Årdal, J. Guzman, D. Edwards, S.A. Matlin and R.F. Terry (2013): «Mapping of available health research and development data: what’s there, what’s missing, and what role is there for a global observatory?», The Lancet, 382(9900).

WHO Global Health I+D Observatory: http://www.who.int/research-observatory/en/

WHO Global Health Estimates: http://www.who.int/healthinfo/global_burden_disease/en/