Appendix B: Additional Data and Analysis
[ Table of Contents ](A) Funding Sources for Canadian PELC Research
We were not able to do an exhaustive enquiry of potential funding sources, but obtained some information through an analysis of funding sources acknowledged in recent PELC publications (Table 1). Notably, almost half of the papers did not acknowledge any funding sources, though individuals and organizations were thanked for assistance. Most of these publications were reviews, guidelines, case reports, anecdotal or editorial in nature and would reasonably not require research funding, but there were also original clinical research studies and an RCT that appeared to have been conducted without external funding. (It is also possible that some authors omitted to acknowledge funding sources, but this is unlikely, since such an omission is considered unacceptable by funders nowadays).
| Table 1 Funding Sources for Canadian PELC Research: sample of 75 publications from 2005-09 where full-text versions could be analysed. | ||
|---|---|---|
| Publications with no funding sources acknowledged: | 34 | |
| Where funding source acknowledged: | 41 | |
| Federal government | 26 | 36 % |
| CIHR | 20 | |
| Canadian Health Services Research Foundation | 3 | |
| Health Canada/PHAC | 2 | |
| NSERC | 1 | |
| Provincial government | 10 | 14 % |
| provincial MOH or equivalent | 7 | |
| provincial government health research agency | 3 | |
| Institutional | 14 | 19 % |
| local health authority/hospital/cancer agency | 9 | |
| university | 4 | |
| professional college | 1 | |
| Health charity | 18 | 25 % |
| local cancer foundation/charity | 7 | |
| Canadian Cancer Society/NCIC | 5 | |
| other health charities | 6 | |
| Foreign | 4 | 5 % |
| National Institutes of Health (USA) | 3 | |
| Veteran's Admin. (USA) | 1 | |
| Private sector | 1 | 1 % |
| Multinational pharmaceutical co. | 1 | |
| Total acknowledgements in 40 papers | 73 | |
Among the studies that acknowledged a funder, CIHR was the most frequently cited individual funder, both for operating and personnel support. Organizations and agencies with responsibilities for health care services and policy collectively accounted for about 25% of the acknowledgements, revealing a strong interest from the users of health research. Charitable sector support related mostly to cancer palliation. Support from industry was notable by its absence.
In order to obtain an approximation for total funding of PELC research in Canada, we can perform some shaky extrapolations from this limited sample. Knowing CIHR's spend on PELC research and the number of acknowledgements this yields, we can calculate the funding that yielded all the acknowledgements, making the breathtaking assumption that each acknowledgement is for the same amount of funding. This yields a total funding for PELC research in Canada of $13M/year during the period 2005-08.
No respondents found that any additional sources of funding had opened up as a result of the increased awareness or activity generated by the CIHR initiative. However, respondents did note that as a result of the initiative, substantial funding was now available to do research, and in addition, the more the overall quality of Canadian PELC research improved, the more researchers were able to access existing (non-PELC specific) funding sources.
One NET, which had developed a strong interdisciplinary team incorporating existing clinical PELC researchers with new social sciences component, was able to win a large Social Sciences and Humanities Research Council (SSHRC) grant for which the original PELC researchers would not have been eligible. (However, this door has now closed for others since SSHRC no longer accepts applications related to any area of health research.)
(B) Productivity of NET PIs
Prior to receiving NET funding, in 2001-03, the future NET PIs as a group were authors on 29% of the Canadian PELC publications appearing during this period. In 2006-08, after receiving NET funding, those same investigators had increased their productivity from 85 to 187 publications over the three-year period, and were authors on 37% of Canadian publications.
| Table 2 Productivity of NET PIs | |||
|---|---|---|---|
| 2001-03 | papers | authors | authors/paper |
| Future NET PIs | 85 | 351 | 4.1 |
| Remainder of Canadian PIs | 209 | 848 | 4.1 |
| % authored by future NET PIs | 29% | ||
| 2006-08 | papers | authors | authors/paper |
| NET PIs | 187 | 1053 | 5.6 |
| Remainder of Canadian PIs | 320 | 1173 | 3.7 |
| % authored by NET PIs | 37% | ||
(C) Canadian PELC Citations
| Table 3: Canada PELC citations/publication relative to World | |||
|---|---|---|---|
| Average citations/publication | Canada | World | Canada/World |
| 1997-09 | 25.5 | 14.2 | 1.79 |
| 2001-03 | 13.0 | 8.8 | 1.48 |
| 2006-08 | 2.35 | 2.06 | 1.14 |
Citation analysis of publications is commonly used to assess the impact of an individual's or a nation's contributions to a field of science, on the grounds that publications that are cited most often by other researchers are those that have the greatest impact on the conceptual or technological development of a field. This is a very narrow approach to the question of impact, since it is equally likely to count as influential a publication that is cited because it is egregiously faulty, or ignore a publication that had huge real-world impact by being the basis for new clinical guidelines, but the guidelines themselves may be disseminated in a form not recognized by the indexing services. However, it is one of the few metrics for assessing research impact, and so we will use it. Table 3 shows the number of citations received per publication, again with data pooled over three-year periods because the small number of Canadian publications resulted in considerable annual fluctuations in citations/publication. The number of citations/publication decreases in more recent periods for both Canada and the world, not because of declining quality of research, but because newer publications haven't had as much opportunity to be cited by other authors. Importantly, publications from Canada in the field of "palliative care" are cited, on average, more often than the world average, suggesting that Canadian work in this field has a relatively high impact. The declining Canada/World ratio over time suggests that other nations are also publishing recent papers with higher impact.
It is tempting, but risky, to try to compare citations/publication for Canadian publications in PELC with other health research disciplines to try to answer the questions "How good is PELC compared to other health research disciplines?" The problem is that different disciplines have different publication cultures. For example, biochemists publish lots of papers, and there are lots of biochemists, so citations/paper are high in biochemistry. Social scientists, on the other hand, may publish mostly books, and correspondingly infrequently. So the citations/publication for biochemistry are much higher than for sociology, but this does not mean that biochemistry in Canada is "better" than sociology. With this big reservation in mind, in Table 4 we have compared PELC publications with those of roughly comparable clinical disciplines. These data are conveniently calculated and presented in the SCImago1 synthesis of the Scopus database, but are not yet available for 2008. Therefore, we have simply compared two time periods, 1997-09 and 2005-07.
| Table 4: Publications in PELC and other health disciplines | |||||
|---|---|---|---|---|---|
| Type | Palliative care | Nursing | Health professions* | Psychology | Medicine |
| 1997-9 | |||||
| Publications | 258 | 575 | 250 | 1859 | 27567 |
| Citations | 6580 | 10361 | 4833 | 30207 | 683579 |
| Citations/Pub. | 25.50 | 18.02 | 19.33 | 16.25 | 24.80 |
| % international authors | 19.0 | 30.0 | 25.6 | 29.9 | 29.5 |
| world share % | 6.1 | 3.9 | 4.1 | 5.6 | 2.8 |
| 2005-07 | |||||
| Publications | 314 | 1237 | 342 | 2588 | 46486 |
| Citations | 1153 | 3338 | 749 | 4763 | 155251 |
| Citations/Pub. | 3.67 | 2.70 | 2.19 | 1.84 | 3.34 |
| % international | 37.0 | 38.3 | 48.9 | 55.3 | 41.9 |
| world share % | 6.0 | 4.7 | 3.9 | 6.0 | 3.3 |
(D) Canada's most productive PELC authors
One remarkable feature of the publishing Canadian PELC research community has been the turnover of the most-productive authors, and their increased productivity. Figure B1 shows the publication history for the ten most productive Canadian authors in 1996-98, in terms of their ranking. For example, the third-most productive author (triangles in Figure B1) in 1996-98 was the eighth-ranked in 2001-03, and 22nd in 2006-08. Only one of the original 'top ten' authors has remained in this august company throughout the analysis period (red line in Fig B1 and B2). We must note that to be included in the top ten today requires considerably more productivity: in 1996-98, five publications/three years earned a spot in the top ten: ten years later this productivity would put an author in 33rd position, and eight publications/three years were necessary to earn a top ten position.
The loss in position of the most productive 1996-08 authors is not simply due to their maintaining their productivity while newer authors leapfrog past them: collectively, the top 10 authors in 1996-98 had 109 publications, but only 47 in 2006-08. Two of them were not represented at all among the 2006-08 authors.
If we examine the publication history of the 2006-08 top ten authors (Fig B2), we see the converse: that is, nine were not among the top ten of 1996-98 (note vertical scale difference between Figs B1 and B2), and only three were among the top ten of 2001-03. Five of the ten were absent from among the 2001-03 authors. In other words, the current most productive authors have arrived recently, replacing the earlier leaders who have retired from active research.
(E) International collaboration
Data on international collaboration in the field of PELC were more easily obtained from the Thomson-Reuter Web of Science2 (WoS), where they are available as a standard report. Since, as noted previously, there were fluctuations in WoS data, and because the number of publications in each year is rather small, in this and other analyses we have compared three separate three-year periods: 1996-98 (MRC, pre-CIHR era); 2001-03 (early CIHR, pre-Initiative era); and 2006-08 (Initiative era). Fig. B3 shows that international collaboration in this field was very low (compared to other Canadian clinical research disciplines (Table 4) during the pre-CIHR era, and increased only slightly in 2001-03, but has risen sharply into the latest period so that almost 40% of publications have at least one international co-author, a level comparable with that of Canadian clinical research discipline publications as a whole (Table 4)). Thus the PELC research community has transformed itself from one that engaged in very little international collaboration to one that is highly collaborative. Fig B3 shows the nationality of the major collaborating co-authors as the % of Canadian papers with a co-author from the specific country, with a threshold of 1%. Collaboration increased most significantly with the US during the early years of CIHR, and more recently with England and Australia. Collaboration with France (and other European nations) has decreased in relative, but not absolute terms.
Fig B3 International Collaboration in PELC Research
(F) Canadian research specializations
Fig B4 Index of Specialization, Health Disciplines
The current Canadian 8% world share of publications in this field is much higher than Canada's overall world share of health research publications (4-5%), showing that Canada specializes in PELC research. The relative degree of specialization in PELC and selected other fields of health research is shown in Fig B4 for the period 1996-2007: an index of specialization =1 means that Canada's share of world publications in the field is the same as the average of all fields. Together with the related fields of geriatrics and gerontology, PELC is one of the areas of health research in which Canada is most specialized.