The priority attached to clinical services for those with respiratory illness varies from country to country. Some such as Australia and Finland have national initiatives for asthma,¹ while others such as the UK have priority arrangements for cancer, including lung cancer. The World Health Organization (WHO) has proposed strategies for the prevention and control of chronic respiratory disorders.^2,3 However, it is unclear whether research spending follows clinical priorities and whether such spending parallels the burden of disease in individual countries. We have sought to analyse the outputs of research papers in respiratory medicine from around the world, to compare them with the burden of disease in individual countries and, for UK papers, to determine the sources of funding for the research.

It is debatable whether research activity should reflect the relative disease burden, either globally or within a country. Some basic research may also benefit several disease areas. Nevertheless, there has been increasing discussion as to whether research budgets should take account of the relative burden of disease and the numbers of patients affected. In the USA increased financial resources given to the National Institutes of Health have led patient advocacy groups to demand a more “equitable” share of research grants.⁴ Globally, there are gaps between the resources devoted to diseases that affect the rich and the poor.^5–9 Recent debate has concerned whether greater funding of cell and molecular biology has delivered what was promised to respiratory medicine.¹⁰ Such debate can only be meaningfully serviced by better understanding of what determines research priorities and by having available data on the distribution and funding of current research outputs.

METHODS

Articles and reviews related to respiratory medicine for the years 1996–2001 were identified and downloaded from the Science Citation Index (SCI) (Institute for Science Information (ISI)) by means of a specially designed “filter” based on the names of specialist journals and title keywords.¹¹ The filter was estimated to have a specificity (or precision) of 0.97 and a sensitivity (or recall) of 0.89, so that it gave a much better coverage of the sub-field than reliance on journal sets. Papers from 15 selected OECD countries were then identified on the basis of their addresses, all of which are recorded in the SCI. These countries, together with their digraph ISO codes used in the tables and figures, are shown in table 1.

Each downloaded paper was then allocated to one of 14 subject areas listed in table 2, with trigraph (three letter) codes used also in table 3 by means of a similar process. As an example of the process, title words or parts of words used by themselves to identify papers in “apnoea and other sleep related breathing disorders” included: apnea, apneic, apnoe*, apnoeic, daytime sleepiness, Epworth, hypopnea, hypopnoea, polysomnographic, snorer, snoring and somnolen*. (The suffix asterisk implied that anything following the asterisk would be included if the preceding letters were present.) Title words or parts of words occurring in combination were also used to identify papers. Examples from the same subject area were: breathing+sleep, hypoventilation+obesity, obstruct*+sleep, sleep+ventilat*, sleep+upper airway, and CPAP + (compliance or sleep or therapy or treatment).

• Download figure
• Open in new tab
• Download powerpoint

Table 3

Relative commitments of 15 OECD countries to different subject areas (other than miscellaneous) within respiratory medicine research in relation to their relative presence in biomedical research. Values <0.25 are shaded dark; values 0.25–0.5 are lightly shaded; values >2 are shown in bold; values >4 are in bold and boxed.

A similar scheme was used to allocate papers to the other subject areas; for some, specialist journals were also used. Any paper that could not be allocated to one of the 13 defined subject areas was classified as miscellaneous; this classification clearly contained many papers reporting research in basic mechanisms.

Each country’s relative commitment to research in respiratory medicine overall and to the individual subject areas was calculated as the ratio between its percentage presence in the subject and its percentage presence in biomedical research overall.^12–14 Thus, Canada publishes 4.7% of biomedical research overall but 5.6% of research in respiratory medicine; its relative commitment to the subject is therefore 5.6/4.7 = 1.19 (the percentage figures have been rounded, so the above calculation is not exact). These relative commitments were then compared with the known death rates per 100 000 population (averages for males and females) for those disease areas where it was possible to obtain data from WHO national mortality tables¹⁵ by means of scattergraphs. For asthma the comparison was with childhood prevalence taken from the ISAAC study¹⁶ and, for cystic fibrosis, with the prevalence of the disease at birth.^17,18

The UK papers were matched with papers in the Research Outputs Database¹² which lists all UK biomedical papers whose funding acknowledgements have been inspected in libraries to determine their sources of support. These were grouped into main sectors as follows:

• UK government, including departments and agencies such as the Medical Research Council, Department of Health, and Biotechnology and Biological Sciences Research Council.

• UK private-non-profit, including charities such as the British Lung Foundation, National Asthma Campaign, Cystic Fibrosis Trust, Cancer Research UK, and foundations such as the Wellcome Trust.

• Industry – mainly pharmaceutical companies.

• International, including the European Commission, the WHO and the US National Institutes of Health.

RESULTS

World papers

A total of 81 419 respiratory medicine research papers were identified between 1996 and 2001. Table 1 shows the mean annual outputs for the 15 OECD countries used for the analysis with, for comparison, their annual percentage presences in respiratory medicine and in biomedicine, and the ratio between them. Overall, Finland, Canada, Spain, and the UK are seen to be relatively strong in respiratory medicine (fig 1).

The breakdown of publications in each of the 14 subject areas is shown in table 2 for two 3 year periods. (The subject area “bronchitis (except chronic)” proved to be so small that no further analysis was applied to this group of papers.) After “miscellaneous”, asthma, lung cancer and paediatric lung diseases accounted for the three largest numbers of papers.

For each of the respiratory subject areas we calculated a country’s relative commitment to that subject area as a ratio of its presence in biomedicine. The results are shown in table 3. Australia, Canada, Finland and Sweden are seen to have a strong publication presence in sleep related breathing disorders. Sweden, Finland, Australia and the UK are well represented in research publications related to asthma. Australia and the UK appear to have a strong commitment to cystic fibrosis research, and the Netherlands a very strong commitment to chronic obstructive pulmonary disease. France and the UK are relatively well represented in the small number of research publications output related to diffuse interstitial lung disease, while Spain is particularly strong on research in the pulmonary manifestations of HIV/AIDS. Australia, the UK, the Netherlands, the USA and Finland are major producers of research publications in viral infections, including influenza (but excluding HIV). Finland dominates research outputs in occupational lung diseases and in paediatric pulmonary research (with Australia) and pneumonia (with Spain). Spain also has a significant research publication output in the field of tuberculosis.

We found relating research outputs to the burden of the disease for each country to be difficult, especially for disease areas such as sleep related breathing disorders where there are very poor records of the burden of disease. In most subject areas correlations were weak. As an example, fig 2 shows the relative commitment to respiratory cancer research related to death rate. Italy and Finland can be seen to do more than the mean for the other countries. The UK output is quite low, perhaps especially in view of its high lung cancer mortality rates. Correlations between research commitment and death rates were effectively zero for tuberculosis and for asthma and COPD, although there was a weak positive correlation with asthma prevalence (data not shown). For publications on pulmonary complications of HIV and AIDS there was a positive correlation with mortality (fig 3), with Spain having the highest death rate (among the countries studied) and the highest research commitment. The USA has a high death rate but only a modest research publication output in this subject area. For cystic fibrosis the UK and Australia had the highest relative commitment but also a high incidence of the disease. Finland has the lowest output and the lowest incidence of the disease (fig 4).

Funding and geographical distribution of UK research

Declared sources of sponsorship and funding for UK research papers published in the years 1996–2000 were analysed. Of 7401 papers, 7048 (95%) had been inspected and 4307 of them (61%) had one or more funding acknowledgements. This is lower than for total biomedical research overall in the UK.¹² The leading sources of declared research sponsorship are shown in table 4.

Figure 5 shows that the amount of stated research sponsorship varies by disease area, with influenza followed by tuberculosis, asthma, pneumonia and HIV/AIDS having the most declared support and pulmonary circulatory disease and diffuse parenchymal lung disease having the least. There were also variations in the amount of support from different sources. The highest levels of government support are in influenza (50%), tuberculosis (39%), and occupational lung diseases (36%); charities and foundations favour cystic fibrosis (35%) and respiratory cancer (34%); and the pharmaceutical industry supports primarily asthma (37%), COPD, and influenza (28%). The percentages of support from the individual sectors, together with those papers without financial acknowledgements, add up to more than 100% because many papers have funding from several sectors in parallel.

DISCUSSION

This report describes an observational tool, use of which has provided an interesting insight into the relative volumes of published research output on a variety of respiratory conditions in different countries. Most countries appear to commit to research into respiratory disease at approximately the same relative level as in biomedical research overall, although Finland and Canada, for example, seem to do rather more respiratory research and Germany rather less. Different countries will be interested in different aspects but, in the UK, it can be seen that lung cancer is relatively under-researched considering the very high burden of the disease in this country. On the other hand, cystic fibrosis research is particularly strong relative to the clinical burden of the disease, and a sizeable part of the funding comes from the Cystic Fibrosis Trust, a charity which is the acknowledged source of sponsorship on 99 out of 379 UK papers (26%). Another apparent research strength in the UK is asthma where there is a relative commitment 1.7 times that in biomedicine overall. Again the charities and foundations play a substantial part, as does the pharmaceutical industry, with one large company (GlaxoWellcome* plc) acknowledged on 13% of the UK output (205 of 1595 papers inspected). This company also funded 14% of UK influenza research and 11% (23 out of 216 inspected papers) in COPD. Of the 431 inspected papers concerning UK respiratory cancer research, the charity Cancer Research UK† supported over 22% of the UK output. Government support for cancer research was rather low at 19%, being proportionately less than its support for influenza, tuberculosis, pneumonia, HIV and occupational lung diseases.

Some disease areas are undersupported in terms of both government research spending and sponsorship by charities, foundations, and the pharmaceutical industry. For example, it may be seen that, whereas research into infectious diseases is rarely published without acknowledgement of a funding source, the majority of research publications related to apnoea and sleep related breathing disorders, pulmonary circulatory disorders, and diffuse parenchymal lung disease have no outside stated sponsorship. There may be a number of explanations for this. These may be areas that are therapeutically unattractive to major pharmaceutical company sponsors or they may be areas in which there are no disease-specific charities. They may represent areas in which there is an absence of data regarding the burden of disease and the need for research may therefore not have been appreciated by funding bodies. Nevertheless, these are important areas—see, for example, recent work on sleep apnoea.^19,20

Research publication outputs may not necessarily accurately reflect all work being done in that area. For example, work done by the pharmaceutical industry may not be published for reasons of commercial confidentiality. Outputs may similarly not necessarily equate to quality, but a cursory glance at Citation Indices does suggest an association between output and the more frequently cited journals, but further work may be needed to elucidate this.

The availability of data is an important starting point for future strategic planning. This observational study of international research outputs in different disease areas has shown that there are variations in the importance attached to respiratory research in different countries, and that that importance does not always relate to the burden of disease in those countries. In some countries the burden of disease—for example, lung cancer in the UK—would suggest that a greater effort is needed into research in that area. It is also important for the sponsors of research (both governmental and private-non-profit) to be aware that “new” areas such as sleep related breathing disorders and those without the likelihood of attracting pharmaceutical company funding should not be overlooked.