APPENDIX 15
Memorandum from City University, Centre
for Information Behaviour and the Evaluation of Research
BACKGROUND
City University has, for the last three years,
maintained and developed on behalf of the Wellcome Trust the UK
Research Outputs Database (ROD) of biomedical research papers.
This now covers 14 years, from 1988-2001, and over half a million
papers nearly all of which have been looked up in libraries so
that their funding acknowledgements can be recorded. [The purpose
of this exercise was to create lists of papers acknowledging the
support of the Wellcome Trust, and of other UK funders of biomedical
research such as the Medical Research Council and medical charities.]
This has enabled us to learn a great deal about how UK biomedical
research is funded. We have also been able to determine the characteristics
of such funding and its influence on the reception accorded the
resulting papers when they are submitted to different journals.
This is likely to be important in any discussion of the possible
change from a system in which scientific journals are paid for
primarily by subscribers (both libraries and individuals) to one
in which the costs are met by means of "page charges"
levied on some or all authors. The latter system might allow the
full text of papers in the journals to be made available gratis
over the web, so benefiting researchers, particularly ones in
low-income countries.
FUNDING ACKNOWLEDGEMENTS
Surveys of researchers have shown that most
are aware of the need to reflect their funding sources accurately
in their acknowledgements. This is not only courteous but is,
increasingly, mandatory for papers in many biomedical journals
because of the need to declare any possible conflicts of interest.
In 1995, we found that almost 90% of funding acknowledgements
were actually recorded and the percentage has probably risen since
then. Nevertheless, about 35% of the papers recorded in the ROD
have no financial acknowledgements, either explicit or implicit
from their corporate addresses. [Papers published by pharmaceutical
companies, or from research council laboratories, would not usually
bear an explicit acknowledgement of support from the company or
research council.] Such papers are sometimes described as "self-funded";
in practice they would normally stem from NHS hospitals or from
academic staff in universities working in their own time. It is
likely that, in the absence of research support from a sponsor,
it would have been difficult for the authors of such papers to
publish them, had they been required to pay page charges; this
could affect about one third of all UK biomedical papers.
What are the characteristics of these papers?
First, they are primarily concerned with subject areas of direct
relevance to the treatment of patients, such as surgery (74% of
papers in this subject from 1989-2000 had no funding acknowledgements),
otorhinolaryngology (66%) and dentistry (64%). Second, within
a given subject area such as cancer research, the unfunded papers
come more from hospitals (41% of NHS papers in 1988-98 have no
acknowledgements) than from universities (28%). Third, the absence
of funding acknowledgements occurs much more on papers in journals
reporting clinical work than in journals reporting basic research.
Figure 1 shows the percentages of unfunded papers for journals
categorized by their research level, on a four-point scale devised
by CHI Research Inc.

Figure 1
Propensity of UK cancer research papers, 1988-98,
in the ROD to carry funding acknowledgments by research level
of the journal in which they are published.
Similar results have been observed in many other
biomedical subject areas. It therefore seems likely that a move
to a page-charge environment would bear disproportionately heavily
on clinical research.
COST OF
RESEARCH PAPERS
If page charges are to be levied on authors,
they may amount to a small or large percentage of the costs of
the research, depending upon the length of the paper and the "cost"
of its preparation. The latter is almost impossible to determine
in individual cases, but an average can be found if leading researchers
are asked about their budgets and these are compared with the
numbers of papers they have written, with account taken of the
fractional contributions of collaborating laboratories. We have
carried out such an exercise [for the Global Forum for Health
Research in Geneva] in eight biomedical research areas based on
particular diseases: cardiovascular, dengue, diabetes, HIV/AIDS,
lower respiratory tract infections, malaria, mental health disorders
and tuberculosis. Internationally, the cost per paper varied from
£83,000 in diabetes (based on 2001-02 US dollars, converted
to pounds at $1.81 = £1) to £234,000 in tuberculosis,
where the high cost of clinical trials on new drugs needed to
combat drug-resistant strains was apparent. There was some evidence
that the costs per paper were higher in high-income countries
such as the USA and Switzerland than in the UK. They would, of
course, have been much lower in low-income countries but there
were few respondents to our surveys in these countries.
A further issue concerns the attribution of
page charges to individual research sponsors. UK biomedical papers
with only one acknowledged source of funding represented 54% of
the total of "funded"" papers in 1988-89 but only
41% in 1999-2000, when more than 8% of these "funded""
papers acknowledged five or more sponsors. Many of them were from
overseas and may or may not be willing to bear their share of
the additional research costs, even if it can be fairly apportioned.
Within the UK, it is the policy of some medical research charities
to give small grants in the hope that their grantees will go on
to attract more substantial funding from bodies such as the Medical
Research Council and the Wellcome Trust, once they have demonstrated
some success. The result is that they are often able to obtain
"gearing" for their money, and be associated with many
research publications to whose cost they have contributed only
modestly. One of the consequences of a move to a page-charge environment
is that less well-funded participants in a research project may
find it harder to be accepted as joint authors of a research paper.
January 2004
|