Scientific production in health-related disciplines in WHO s Western Pacific Region in

Transcription

1

2

3 WHO Library Cataloguing in Publication Data Scientific production in health-related disciplines in WHO s Western Pacific Region in Health services research. 2. Scientific literature. ISBN (NLM Classification: WA 84.3) World Health Organization 2009 All rights reserved. The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement. The mention of specific companies or of certain manufacturers products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters. The World Health Organization does not warrant that the information contained in this publication is complete and correct and shall not be liable for any damages incurred as a result of its use. Publications of the World Health Organization can be obtained from WHO Press, World Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel: ; fax: ; bookorders@who.int). Requests for permission to reproduce WHO publications, in part or in whole, or to translate them whether for sale or for noncommercial distribution should be addressed to WHO Press, at the above address (fax: ; permissions@who.int). For WHO Western Pacific Regional Publications, request for permission to reproduce should be addressed to Publications Office, World Health Organization, Regional Office for the Western Pacific, P.O. Box 2932, 0, Manila, Philippines, fax: , publications@wpro.who.int

4 TABLE OF CONTENTS ACKNOWLEDGEMENTS EXECUTIVE SUMMARY 1. METHODOLOGICAL CONSIDERATIONS 1 2. PRODUCTION OF SCIENTIFIC ARTICLES IN THE WESTERN PACIFIC REGION 3 3. COLLABORATION PATTERNS Collaboration of the Western Pacific Region with other WHO regions Collaboration by countries income group Collaborations by group of countries VISIBILITY OF THE PRODUCTION OF SCIENTIFIC ARTICLES IN THE WESTERN PACIFIC REGION Production visibility by WHO regions Visibility and collaboration in the Western Pacific Region PUBLIC HEALTH RESEARCH PRODUCTION IN THE WESTERN PACIFIC REGION 23 STATISTICAL ANNEX 26

5 ACKNOWLEDGEMENTS This report was prepared by Dr Guillermo Paraje from Universidad Adolfo Ibáñez in Santiago, Chile, building on analysis previously conducted in collaboration with Dr Ritu Sadana and the Health Research Systems Analysis Initiative, WHO, Geneva. Dr Reijo Salmela from the WHO Regional Office for the Western Pacific provided regional insights into planning of this study and its findings. WHO is grateful to Thomson Scientific for logistic support in setting up the databases and the contractual agreement to use the databases to support analysis across WHO regions.

6 EXECUTIVE SUMMARY In 2002, the Department of Research Policy & Cooperation at WHO Headquarters in Geneva initiated an intercountry project for health research systems analysis. Since 2006, the WHO Regional Office for the Western Pacific has conducted similar work in selected countries. This report describes trends in scientific production in the Western Pacific Region from 1992 to 2006, with focus on the first 10 years. The first chapter of this report examines the study s methodology. More than 3.5 million bibliographic references in Thomson ISI Web of Science (health-related articles, notes and reviews) were analysed (period ). A broad definition of health was used, covering related social, medical, environmental and physical sciences. Citations received by each reference over a three-year period after publication were also considered. The second chapter reports on the scientific research output of countries in the Region. Among the six WHO regions, the Western Pacific Region was the fastest growing region in health research, increasing its share of global research production by more than 30% from 1992 to Not all countries in the region showed, however, the same level of performance. Of the five largest producers of research (Australia, China, Japan, New Zealand and the Republic of Korea), China and, particularly, the Republic of Korea, showed the most impressive performances (the latter with a ninefold increase in its global share). The third chapter takes a closer look at collaboration patterns. International collaboration meant, with few exceptions, collaboration with countries from other WHO regions, mainly with the United States of America and high-income European countries. This type of collaboration represented around 66% to 90% of all international collaboration. Australia and Japan, high-income countries, collaborated extensively and almost exclusively with countries outside the Region. More than 80% of their iternational collaboration was conducted in this way, whereas only around 10% of their international collaboration was done with other Western Pacific Region countries. The other three largest producers collaborated more within the region. Around 25% of the papers produced with international collaboration by China and the Republic of Korea involved other Western Pacific Region countries, and in the case of New Zealand (during 2001) intra-regional collaboration reached 35%. Whereas collaboration outside the region was stable over time for Australia and China, it increased in the Republic of Korea and decreased in Japan and New Zealand. Finally, the fourth chapter examines the visibility of the scientific research. Collaboration between low-income or middle-income countries was found mostly in low visibility research. Visibility tends to be higher when collaboration is with countries outside the region. In almost all cases the proportion of high visibility research conducted in collaboration with countries outside the Westrern Pacific Region doubles or more than doubles the low visibility research.

7

8 1. METHODOLOGICAL CONSIDERATIONS Pattern analyses of scientific output reveal not only specific results obtained during the research process, but also countries prospects in adopting new technologies and intentions towards improving the quality of life of their populations. The most used production indicators arise from scientific publications, e.g. scientific journals and congresses, and are relatively easy to measure especially those from scientific publication databases. Since these databases detail where publications were generated, they provide pertinent information on scientific activity in specific institutions or countries. However, they can overlook such indications as unpublished production or training achieved in the production process. Therefore, publication analyses do not exhaust the process of scientific output but do consider an important and easily measurable dimension. The database used to conduct this study was compiled by Thomson Institute of Scientific Information (ISI) and includes three of its products: the Science Citation Index, Social Science Citation Index and Arts and Humanities Citation Index. Although these indexes include various categories of publications, this report will consider only articles, notes and outlines that included a peer review process. These consist of publications in regularly published scientific journals, in which the principal criterion for inclusion was its high impact factor. It is known that the Thomson ISI database has limitations that are shared with most bibliographic databases, such as the exclusion of presentations in congresses and posters. Perhaps most important is that the database is biased towards publications in English and those that originate in Anglo-Saxon countries. 1 However, this database contains the complete affiliation of all publication authors in contrast to MedLine, which only contains information for the first author and contains information on publications bibliographic references, allowing the analysis of the number of citations received after publication. Scientific journals included in the Thomson ISI database are classified in at least one category or scientific field. For example, the Japanese Journal of Physical Fitness and Sports Medicine is classified under physiology and sport sciences. Since the database does not discern health publications, however, it is necessary to make that separation individually. In this report, the definition of health is very broad and includes traditional health fields such as cardiology and paediatrics as well as other fields linked to the social sciences in which scientific journals tend to publish health-related articles, such as in economy or demography. Scientific fields were, in turn, grouped into seven areas used in the forthcoming analyses: clinical medicine, biomedical sciences, basic sciences, pharmacology, food sciences, social sciences and public health. Table 1 shows the grouping of the fields considered in the definition of health based on their primary scientific fields. 1 It is difficult, however, to establish how biased this is as no publication estimates comply with requirements similar to those established by Thomson ISI (e.g. in terms of periodicity of the publication or quality). 1

9 Table 1. Classification of health science fields Although the Thomson ISI database supplies information on the institutional affiliations of all authors, knowing their actual contribution to a publication is impossible. Nor is it possible to assign an ad hoc structure to their contribution since institutional affiliations are not linked to each author. Due to this limitation, it is presumed in this paper that all authors contributed equally to the article. Therefore, output (i.e. an article) is divided into equal parts among all participating institutions. For example, an article in which two Australian, one Fijian and one Japanese institution are credited will mean a contribution of one half of an article for Australia, one fourth of an article for Fiji and another fourth for Japan. This form of counting is arbitrary but has the advantage of simplicity for computing and not distorting the regional totals of articles produced. Finally, much effort has been made to provide data for A web-based search was implemented utilizing Thomson ISI s Web of Science, but data obtained in this manner differ from data for First, the definition of health was different, as the web search was based on a string of keywords and not on the journal classification provided by ISI. The web-based search was therefore narrower, strictly based on traditional health topics. Secondly, the counting method was different. Research production during was estimated using the aforementioned fractional counting method, but for the web searches, a whole counting method was used, which gave one point to each participating country in a publication. Trends obtained using this methodology should be treated as very rough indicatives of what countries and areas produced in and should be used with caution. Such problems arising from web-based searching makes analysis based on collaboration and visibility of health-related research very difficult. 2

10 2. PRODUCTION OF SCIENTIFIC ARTICLES IN THE WESTERN PACIFIC REGION 2 Global health publications, i.e. articles, notes and reviews in scientific journals compiled by Thomson ISI, have evolved in the analysed period ( ). At the beginning of the period, some publications on health were produced, while during 2001, the publications totalled The average annual growth rate of health publications was 1.75%. 3 This growth was unequally distributed in World Health Organization (WHO) regions, as shown in Table 2. The Western Pacific Region grew at an average annual rate of 4.9% the highest rate among all regions. Other fast-growing regions included the Eastern Mediterranean Region with a rate of 4.2%, and the South-East Asia Region with 2.3%. Only the African Region had a negative annual growth rate. As is expected, within regions, countries or groups of countries performed differently. Within the Western Pacific Region, the five largest producers (Australia, China, Japan, the Republic of Korea and New Zealand) grew at an annual rate of 4.9%, whereas the Region s other countries and areas had faster rates, averaging 5.5%. Table 2. Annual average growth rate for health-related publication by WHO regions Regions Annual average growth rate Region of the Americas 0.59 African Region European Region 1.97 Eastern Mediterranean Region 4.18 South-East Asia Region 2.34 Western Pacific Region 4.89 Largest five producers 4.88 Rest of Western Pacific Region 5.53 Total 1.75 Despite its moderate growth rate of 0.6%, the Region of the Americas was, on average, the main producer of health articles during the period analysed, as shown in Figure 1 and Table 3. On average, the Region in which the main research producer, the United States of America (USA), is located contributed 42.9% of global health-related production. The second-largest producer was the European Region with an average of 41.1%, and the third-largest producer was the Western Pacific Region, with an average share of 13.3%. Within this Region, the five largest countries contributed an average of 13.0%, leaving the other 24 countries and areas with the remaining 0.3%. This pattern of production makes the Western Pacific Region one of the most concentrated research-producing regions in the world. 4 2 Most of this section analyses correspond to the period, though some analyses are given for the period. 3 To provide a context for this increase, the world s gross domestic product (GDP) increased at an annual average rate of approximately 3.2% from 1992 to 2001, according to the International Monetary Fund. 4 Within regions, each country s contribution was uneven. In general, a strong concentration of scientific output was observed at the international level when a small group of countries in a region create most of the region s production (Paraje G., Sadana R. and Karam G., Increasing international gaps in health-related publications, 2005, Science 308: ). 3

11 Figure 1. Health-related research production, by WHO region AMR: Region of the Americas; AFR: African Region; EUR: European Region; EMR: Eastern Mediterranean Region; SEAR: South-East Asia Region; WPR: Western Pacific Region Table 3. Health-related research production by WHO region WHO Regions Region of the Americas African Region European Region Eastern Mediterranean Region South-East Asia Region Western Pacific Region Average Largest five producers Rest of Western Pacific Region Table 4 shows that the increase in health-related research production in some countries and areas of the Western Pacific Region is also noticeable in the ranking of global producers. The Region has four countries within the top 20 producers when considering the average production: Australia, China, Japan and the Republic of Korea. Table 4. Top 20 health-related producers Average Top 20 producers United States of America United Kingdom Japan Germany France Canada Italy The Netherlands Australia Spain Sweden China Switzerland India Israel Belgium Denmark Finland Brazil Republic of Korea Rest of the world Total

12 Table 5 and Figure 2 show how the scientific output evolved in each region as a proportion of global scientific health-related output. Both demonstrate a panorama similar to growth rates in regional production. The Western Pacific Region showed the most important growth between 1992 and 2001, with almost a 32% increase in its global share. The second-largest growth was registered by the Eastern Mediterranean Region, which increased almost 24% in the same period. The African Region and the Region of the Americas, on the contrary, ended the period with lower participation than the previous period. Figure 2. Regional shares of world s health-related research production, , (1992 = ) WPR EMR SEAR EUR AMR AFR WPR: Western Pacific Region; EMR: Eastern Mediterranean Region; SEAR: South-East Asia Region; EUR: European Region; AMR: The Americas; AFR: African Region Table 5. Health-related research production by WHO region (1992 = ) WHO Regions Average Region of the Americas African Region European Region Eastern Mediterranean Region South-East Asia Region Western Pacific Region Largest five producers Rest of Western Pacific Region As shown in Figures 3 and 4, strong differences in production existed within the Western Pacific Region. For instance, the Republic of Korea experienced a nine-fold increase in its global share of research production, whereas China s was two-fold. Figure 3. Main research producers in the Western Pacific Region (1992 = ) Republic of Korea China Japan, Australia and New Zealand 5

13 The other main producers also demonstrated significant increases as shown in Figure 4. Japan, the Region s largest producer, had a 10-point increase in its global share of production, as did Australia. New Zealand finished with a similar participation rate to Figure 4. Main research producers in the Western Pacific Region (1992 = ) Japan Australia New Zealand Although the dataset available to WHO ends in 2001, there are some indications that the trends in this period continued. Using Thomson ISI s website and searching for articles, notes and reviews on health topics produced by the Region s five main research producers, research production continued to grow strongly. 5 Though these results are not directly comparable with those obtained for period, they demonstrate an important increase in the global share of research production of the five countries since Figure 5. Main research producers in the Western Pacific Region (2001 = ) 5 The data obtained from Thomson ISI s website are different from those obtained for Please see Section 1 for a discussion on webbased searches. 6

14 Figure 5 shows that China and the Republic of Korea continued to lead the growth rates, with an increase around 150% for both in their respective global shares. Australia, Japan and New Zealand also had important increases ranging from 50% to around 25%. Given the regional importance of these five producers, it was likely that the Region s share in world research production increased rapidly during Table 6 displays the evolution of the global share of health-related research production for the Region s 24 countries and areas, 6 which have been grouped into three ad hoc groups. The first group consists of the five largest producers, representing 13% of the world s production on average. Within this group, Japan contributed an average of 8.1% of the world s research production, Australia contributed 2.3%, China 1.6%, the Republic of Korea 0.6% and New Zealand 0.4%. The 20-fold difference between the research share of the first producer (Japan) and the fifth (New Zealand) is a good indicator of the high research concentration that existed in this Region. Table 6. Health-related research production in countries and areas of the Western Pacific Region, percentage of world production Average Countries Five largest producers Australia China Japan Republic of Korea New Zealand Pacific Islands Cook Islands Fiji Marshall Islands Federated States of Micronesia Palau Papua New Guinea Samoa Solomon Islands Tonga Denmark Tuvalu Vanuatu Other countries Brunei Darussalam Cambodia Lao People s Democratic Republic Malaysia Mongolia Philippines Singapore Viet Nam TOTAL The second group consists of Pacific islands, which contributed an average of only % of the world s research production. Even Fiji and Papua New Guinea had very small participation rates. The third group, a heterogeneous group of mid-sized and small countries, contributed an average of 0.26% of world s research production in health-related topics, with Singapore representing more than two thirds of this group s production. Malaysia comes in second with around % of world s production. 6 Only countries or areas that had at least one publication in the period are shown. 7

15 Table 7 shows the average number of papers in which institutions from the Region s countries have been involved. 7 Unsurprisingly, Japan appears here in first place with around 59% of the publications. Australia has 18% of the total, while China has 13%. Table 7. Average annual number of publications by countries of the Western Pacific Region Country Average Share (%) Australia Brunei Darussalam Cambodia China Cook Islands Fiji Japan Republic of Korea Lao People s Democratic Republic Malaysia Marshall Islands Federated States of Micronesia Mongolia New Zealand Palau Papua New Guinea Philippines Samoa Singapore Solomon Islands Tonga Tuvalu 0.0 Vanuatu Viet Nam Total Table 7 explains the modest involvement that many of the Region s countries had in the research production process at least when considering publications indexed by Thomson ISI. Malaysia, the Philippines and Viet Nam, even when jointly considered, did not reach a 1% contribution rate. Figure 6 details the Region s research concentration. Even when excluding the five largest countries, the concentration of research production was very large. Indeed, adding up the contribution of the second five-largest producers Malaysia, Papua New Guinea, the Philippines, Singapore and Viet Nam they showed a similar concentration level of 96%. This fractal pattern of production points towards three levels of countries: the large world producers (Australia, China, Japan, Republic of Korea and New Zealand) that had an important share in world research production and led the Region s increase in the world s research production, the large regional producers (Malaysia, Papua New Guinea, the Philippines, Singapore and Viet Nam) that had limited participation in world s research production but are significant in the Region, and the small producers that had very little participation in the Region s research production. 7 Countries contributions were counted by giving one unit to each participating country in the publication, i.e., the whole counting method The total in Table 7 should not be treated as the total number of papers produced with the involvement of these countries, as some countries may have collaborated on a publication and were counted twice. 8

16 Figure 6. Average annual number of publications in the countries in Western Pacific Region in The research pattern displayed in Table 6 can be also considered within the regional context. Table 8 displays the participation and evolution of countries research production as a share of regional production. Line one of this table demonstrates the extreme concentration of research production in the Region: the five largest countries produced 98% of the health-related publications. Japan, the largest producer, contributed an average of 61.5%, though its regional share fell steadily from 66.0% in 1992 to 56.0% in This decrease was not produced by a decline in the number of publications in which Japanese scientists were involved, but by other countries large increases in production. In other words, Japan s decrease was not in absolute but in relative terms. The second largest producer, Australia, experienced the same pattern, as its participation in regional production decreased from 18.4% in 1992 to 15.4% in

17 Table 8. Health-related research production in countries of the Western Pacific Region, percentage of regional production Countries Average Five largest producers Australia China Japan Republic of Korea New Zealand Pacific Islands Cook Islands Fiji Marshall Islands Federated States of Micronesia Palau Papua New Guinea Samoa Solomon Islands Tonga Tuvalu Vanuatu Other countries Brunei Darussalam Cambodia Lao People s Democratic Republic Malaysia Mongolia Philippines Singapore Viet Nam TOTAL The decline in these countries was due to the rising production rates of China and the Republic of Korea. Their scientific production reflected the upsurge in their economies and aggressive scientific strategies, which were aimed at gaining presence in international journals. China s regional share doubled from 8.8% in 1992 to 16.4% in 2001, while Republic of Korea s participation went from 1.2% in 1992 to 8.0% in Other countries also experienced increases. Singapore s regional share grew from 1.1% in 1992 to 1.4% in In fact, at the beginning of the period, Singapore and the Republic of Korea had similar research production, but the Republic of Korea vastly surpassed Singapore by the period s end. This pattern may be explained by the early development of Singapore s health research system, making Singapore reach its publication peak in the early 1990s, or perhaps by different research policy strategies. Table 9 shows the scientific output for the 10 largest producers in the Region (which comprise more than 99% of the Region s production) by scientific area for Such areas (defined in Chart 1) corresponded to where a scientific journal originated and not to the publication itself. A scientific journal can be classified into several areas; therefore, the total of the scientific output by area surpassed the total scientific output of a country. 10

18 Table 9. Research production by scientific field for the 10 largest research producers, 2001 Number of Publications Rep. of New Papua New Scientific Field (1) Japan Australia China Korea Zealand Singapore Malaysia Philippines Viet Nam Guinea Clinical medicine Biomedical sciences Basic sciences Pharmacology Food sciences Social sciences and social welfare Public health and health systems TOTAL NUMBER OF PUBLICATIONS (2) Number of Publications Rep. of New Papua New Scientific Field (1) Japan Australia China Korea Zealand Singapore Malaysia Philippines Viet Nam Guinea Clinical medicine Biomedical sciences Basic sciences Pharmacology Food sciences Social sciences and social welfare Public health and health systems TOTAL NUMBER OF PUBLICATIONS (2) (1) Scientific fields are defined in Table 1. (2) Since scientific publications may be classified in more than one field, some percentages may be greater than %. Except for the Philippines, all countries in this group participated in publications on clinical medicine the most. In the largest producer of the Region, Japan, this area represented little more than 43%. Biomedical sciences and basic sciences represented the other most important groups. Australia had the highest proportion in biomedical sciences (almost 28% of its production), while China and the Philippines had the lowest percentages with 21% and 19%, respectively. In relative terms, the Republic of Korea was the principal producer in basic sciences, with 35% of its production in this area, while New Zealand and Papua New Guinea produced much less. Pharmacology was the fourth area of importance, with a total production between 5% (Australia) and 26% (Papua New Guinea). For public health and health systems, Japan had an extremely low participation rate at only 1.5%. Japan was not even the main regional producer in this area in absolute numbers, it only produced 481 publications in Among the largest producers, Australia, China and New Zealand had relatively high shares of publications in this area, while Papua New Guinea and Viet Nam (with 17% and 14%, respectively) were also significant contributors. 11

19 3. COLLABORATION PATTERNS The previous section supplies data on the production efforts of countries in the Western Pacific Region and how that production evolved in over a single decade. However, since countries contributions to publications were judged to be equal for simplicity s sake, information on collaboration and potential networks created in that process was lost. Examining how different countries related when publishing the results of their scientific output is useful to discern if collaborations were determined by cultural affinities, geographical proximity or partners economic resources to finance research. In methodological terms, the publications used to record collaboration patterns between regions or countries should be examined differently than the data in the previous section. In the previous section, for example, if a publication had, as scientific authors, two Australian institutions and one Korean institution, for the purpose of analysing the production of both countries, two thirds of the publication is assigned to Australia and one third to the Republic of Korea. However, for this section, what matters is not how much each country contributes but that a collaboration with other countries exists and can be identified. Therefore, in the aforementioned case, it would only be recorded as Australia collaborating with the Republic of Korea (and vice versa). Double accounting of the publications with some international collaboration prevents the participation percentages of each country in the whole Western Pacific Region from totalling %. 3.1 Collaboration of the Western Pacific Region with other WHO regions A first approximation of the Western Pacific Region s patterns of collaboration were demonstrated by how the Region relates to the rest of the world, which is grouped according to the WHO regions. Table 10 displays those patterns. Table 10. Collaboration of the Western Pacific Region with other regions, 2001 AFR AMR EMR EUR SEAR Western Pacific Region, number of publications (1) Western Pacific Region, number of publications with involvement of five largest producers (2) Western Pacific Region, number of publications without involvement of five largest producers (3) Western Pacific Region, with involvement of five largest producers (Percentages of the Region production) (1) Publications in collaboration with the entire Western Pacific Region (2) Publications in collaboration with at least one of the Western Pacific Region s five largest research producers (3) Publications in collaboration with none of the Western Pacific Region s five largest research producers AFR: African Region; AMR: Region of the Americas; EMR: Eastern Mediterranean Region; EUR: European Region; SEAR: South-East Asia Region The Western Pacific Region partnered most with the Region of the Americas and the European Region when publishing scientific articles. During 2001, more than 7200 and 4800 publications, respectively, were written jointly. Of these joint publications, in 97% and 95% of the cases, respectively, there was some involvement from at least one of the Western Pacific Region s largest five producers. The other regions collaborated little with the Western Pacific Region. The South-East Asian Region, the Western Pacific Region s third main partner, only participated in 400 publications; the African Region and the Eastern Mediterranean Region participated in only 174 and 102 publications, respectively. In these publications, large Western Pacific Region producers participated in around 90%. Interestingly, the largest proportional involvement of large producers was with regions with the highest concentrations of high-income countries. 12

20 These figures do not correspond to exclusive bilateral collaborations among the regions, e.g. countries of the Western Pacific Region collaborating only with countries the South-East Asia region, but any kind of collaboration. Therefore, the low numbers obtained for the African Region or the Eastern Mediterranean Region reveal the poor contact of these regions institutions even within the framework of broad collaboration projects including, for instance, European countries. As Table 10 shows, collaboration with some countries from certain regions is sporadic and scarce. What causal factors can explain these collaboration patterns? Unfortunately, the bibliographic databases do not provide direct responses since they only collect information on authors affiliations and place of publication. However, based on these data, it is possible to infer that collaboration was determined, to a great extent, by the existence of economic resources. It was, then, the availability of these resources and not the existence of common research agendas that may have facilitated collaboration between countries or regions. 3.2 Collaboration by countries income group Table 11 shows another aspect of the collaborative relationship between countries in the Western Pacific Region and the rest of the world. Intraregional collaboration is detailed, i.e. research production conducted between at least two institutions from the Region without the involvement of any institution from outside the region, as well as national collaboration, i.e. institutions from the same country participating in the research outcome. The table further demonstrates extraregional collaboration, i.e. research production that includes some participation from countries from other regions, which may feature collaboration with high-income countries, high-middle income countries, low-middle income countries and low-income countries. 8 Thus, for instance, a paper written by a Japanese institution and a Nigerian institution is considered an extraregional collaboration with a low-income country. Furthermore, if an American institution was also involved in that collaboration, then it is considered as an extraregional collaboration with a high-income and a low-income country. 9 Table 11 also supplies information on articles written without any collaboration by a single participating institution. The total number of publications is the sum of intraregional collaborations, extraregional collaborations and no collaborations. Table 11. Collaboration patterns in the Western Pacific Region Type of Collaboration Publications Percentage Publications Percentage Intraregional collaboration (1) Extraregional collaboration (2) with high income countries with high-middle income countries with low-middle income countries with low income countries No collaboration (3) TOTAL (1) Collaboration between at least two institutions from Western Pacific Region (could be from the same country) exclusively (2) Collaboration with at least one institution from other regions (3) Only a single institution involved in the publication 8 This classification of countries based on their average income is sourced from those made by the World Bank for This way of counting publications may cause double-counting; therefore, the total extraregional collaboration is lower than the sum of the collaborations within the different income groups. 13

21 Figure 7 shows the magnitude and direction of the main changes in collaborations that occurred from 1992 to The proportional share of publications without any collaboration (intra- or extraregional) declined: in 1992, little more than 50% of publications were from individual institutions, and in 2001, this figure declined to 35%. The development and extensive use of new remote communication technologies undoubtedly influenced this decline. Figure 7. Collaboration patterns in the Western Pacific Region The relative decrease in noncollaborative research meant an increase in intraregional and extraregional collaboration. The increase in intraregional collaboration was larger (from 36% of total publications in 1992 to 45% in 2001). Extraregional collaboration increased from 13% in 1992 to 19% in This was due to growth in scientific collaboration with high-income countries, which also grew by around 6% (from 12% of total research production in 1992 to 18% in 2001). Overall, collaboration with this group of countries represented around 94% of extraregional collaborations. A closer examination reveals that the USA, the United Kingdom of Great Britain and Northern Ireland, Germany and France were the main collaborators and that they constituted the bulk of extraregional collaboration with the Western Pacific Region. Collaboration with other income groups also increased collaboration with low-middle-income countries more than doubled in this period though its magnitude remained very low. The figures given in Table 11 demonstrate that the availability of resources to finance collaborative project strongly influenced collaborations between countries. This is evident not only by the volume of collaborations with high-income countries in regard to the total collaborations but also the trend followed during the 10 years analysed. 3.3 Collaborations by group of countries A closer look at intrazone collaboration is given in Table 12, showing how the Western Pacific Region subgroups collaborated during The intersections of the table show the number of joint publications each subgroup had exclusively, i.e. only between two of them or as part of collaborations with other subgroups, for intrazone collaboration projects. The figures in bold demonstrate exclusive collaborations within the specific subgroup. 14

22 Table 12. Collaborations among subgroups in the Western Pacific Region, 2001 Largest producers Middle-size producers Pacific islands Largest producers Middle-size producers Pacific islands Intersections for the same subgroup in bold are within-group collaborations, i.e. exclusive. Intersections between subgroups are intraregional collaborations between the two subgroups. Table 12 shows that despite the important increase in intraregional collaborations, collaborations among subgroups was very low. The main collaborations registered are within each subgroup showing, possibly, an important collaboration between institutions of a single country. Collaborations between the largest producers and middle-size producers, i.e. those with small global contributions but a sizeable regional contribution, were just above 8000 publications in The main collaborations were within the largest producers group, reaching almost publications in Table 13 shows the collaboration patterns for the largest producers group for 1992, 1998 and Table Collaboration patterns for the main producers in the Western Pacific Region 1992 Australia China Japan Rep. of Korea New Zealand No collaboration or national collaboration International collaboration - Within Western Pacific Region only - Outside Western Pacific Region exclusively - Within and outside Western Pacific Region TOTAL (Percentage) Total number of publications Australia China Japan Rep. of Korea New Zealand No collaboration or national collaboration International collaboration - Within Western Pacific Region only - Outside Western Pacific Region exclusively - Within and outside Western Pacific Region TOTAL (Percentage) Total number of publications Australia China Japan Rep. of Korea New Zealand No collaboration or national collaboration International collaboration - Within Western Pacific Region only - Outside Western Pacific Region exclusively - Within and outside Western Pacific Region TOTAL (Percentage) Total number of publications Taken from Paraje, S. and Salmela R. Collaboration and visibility of health research in the Western Pacific Region, 2007, presented at the Global Forum 11, Beijing, China. 15

23 Table 13 shows several important characteristics of research collaboration among countries. Well-established, high-income countries, such as Australia, Japan and New Zealand, increased international collaboration. Not surprisingly, relatively small countries, such as New Zealand, were more open to international collaboration possibly because national networks are relatively small and increased such openness over time. A similar strategy is found in other parts of the world, like in Scandinavian countries, and also in larger countries such as Australia and Japan. Australia went from around 19% of total publications written through international collaboration in 1992 to 32% in New Zealand counted 26% in 1992 and 36% in 2001, whereas Japan, 11% and 17%, respectively. Newly-developed countries, such as Republic of Korea, and fast-growing countries, such as China, decreased international collaboration. In 1992, 26% of Chinese publications included some international collaboration, while such a figure descended to 22% in 1998 and then recovered to 25% in The Republic of Korea went from 33% in 1992 to 23% in Unfortunately, the database does not provide further information on the causes of these patterns. It is possible that the return of Chinese students studying abroad produced a sudden strengthening of national networks, which promoted national collaboration. An increasing share of publications in new journals, which are non-isi indexed, is another possibility. The Republic of Korea s case may be related to the introduction of some promotion mechanisms for academics and scientists that may have both increased research production and promoted national collaboration. In any case, the change in percentages has occurred with the stunning increase in research publications, so absolute numbers have grown for all categories. The other characteristic in health research production for this group was that international collaboration means, with few exceptions, collaboration with countries from other regions, mainly the USA and high-income European countries. This type of collaboration represented around 66% to 90% of all international collaborations. Australia and Japan, themselves high-income countries, had higher collaborations with extraregional countries (exclusively). More than 80% of their international collaboration was conducted in this manner, whereas only around 10% of their international collaboration was conducted with other countries of the Western Pacific Region. The other three high-performing countries collaborated more within the Region. Around 25% of papers with international collaboration for China and Republic of Korea involved other countries of the Western Pacific Region, and such a figure reached 35% for New Zealand in Although outside the Region, collaboration was stable for Australia and China, it increased in Republic of Korea and decreased for Japan and New Zealand. 16

24 4. VISIBILITY OF THE PRODUCTION OF SCIENTIFIC ARTICLES IN THE WESTERN PACIFIC REGION While Section 2 is concerned with scientific research output, this section examines the visibility that such production had. This indicator is extracted from the average number of citations that the publications received. As mentioned in Section 1, the Thomson ISI database is the only bibliographic database that includes the references of the publications it compiles. Therefore, it is possible to count which publications are referenced more frequently. This does not indicate a publication s importance or impact, but simply its visibility, since they can be cited critically, i.e. bad publications can, in principle, be cited as much as better ones. Calculating this indicator may be difficult since the number of citations that publications received can depend not only on the type of publication, but also on the scientific area to which they belong. Therefore, a given period of time to count the citations received by each publication in different areas may be arbitrary and, lastly, erroneous. To avoid these issues, the following procedure was adopted. First, the citations received on average by all the publications in a specific scientific area and by type of publication were counted. This was done by counting the publications that appeared in scientific journals that are classified by areas. Since those journals may be classified in various areas, the simple average was taken of the references received by all areas in which the journal was classified. In this way, it was possible to obtain an expected average of references for each type of publication in each scientific area. This average of expected citations was calculated for 3 years after the publication. Hence, for an article published in 1992, the citations received were counted between 1992 and This restricts the analysis period to Once computed, these averages of expected citations for each publication were compared to those effectively received in the same time period. Thus, if a publication received more citations than expected for its type and scientific area(s), it may be said that its visibility is above its peers. This ratio between received and expected citations can range between 0 and 1 (low visibility), between 1 and 2 (high visibility) and greater than 2 (very high visibility). In other words, all publications have an expected citation level CE given by the fields to which these publications belong. Thus, if a determined publication i belongs to a certain field m the number of citations received by all publications J in that field must be averaged: If a publication belongs to more than one scientific field, K, the average of expected citations in each field must be considered and averaged: 17

25 This gives the expected citation for each publication i belonging to these fields. To consider if a publication has had more or less citations than the average publication, a ratio of actual citations to expected citations can be considered. This ratio is defined as: If R i is less than 1, a certain publication k has received fewer citations than the average publication in the same fields and, hence, has lower visibility than average. On the other hand, if R i is greater than 1, it means that such a publication has received more citations than the average publication in the same fields and, hence, has higher visibility than average. 4.1 Production visibility by WHO regions Table 14 shows how scientific output was distributed by WHO region according to its visibility in The first column corresponds to scientific production and is taken from Table 3. The second, third and fourth columns show the distribution of research production, according to its degree of visibility. Thus, the second column shows how much of the low visibility production was produced by each WHO region. For instance, the Western Pacific Region contributed 14% to global research production, but only 9% to global research production with high visibility. Its contribution to global research production of low visibility was 16%. Table 14. Research visibility by WHO region, 1998 Share in global Ratio of expected citations WHO Region research output (%) Lower than 1 Between 1 and 2 Greater than 2 Region of the Americas African Region Eastern Mediterranean Region European Region South-East Asia Region Western Pacific Region Visibility indexes for every region were based on information contained in Table 14 and shown in Table 15. These indexes were calculated as the ratio between the distribution of each region for a given factor of visibility and the regional distribution of production. In this manner, the index corresponding to low visibility is the ratio between column 2 and column 1 of Table 14, that of high visibility is the ratio between columns 3 and 1 and, finally, that of very high visibility is the ratio between columns 4 and 1. An index smaller than 1 means that a given region contributed less to the research production of that level of visibility than to global research production Table 15. Visibility indexes by WHO region, WHO Region Low visibility High visibility Very high visibility Region of the Americas African Region Eastern Mediterranean Region European Region South-East Asia Region Western Pacific Region

26 Table 15 shows that the low visibility production was concentrated mainly in the Eastern Mediterranean Region, South-East Asian Region, the African Region and Western Pacific Region. For example, the Eastern Mediterranean Region had 33% more low visibility publications than the publication total. 11 The South-East Asia Region had a share in low visibility production that surpassed its total participation by 31%, while the African Region and the Western Pacific Region exceeded it by 21% and 10%, respectively. 12 Concerning high and very high visibility, the Western Pacific Region had indexes less than 1, meaning that its share in research production of such visibility was below its average. Therefore, the Western Pacific Region tended to concentrate its research production in low visibility production. Table 16 shows the visibility indexes during 1998 for every country of the Western Pacific Region and for each of the defined subgroups. 13 Table 16. Research production by visibility in countries of the Western Pacific Region, 1998 Subgroups / countries Share in global Ratio of expected citations research output (%) Lower than 1 Between 1 and 2 Greater than 2 Five largest producers Australia China Japan Republic of Korea New Zealand Pacific Islands Fiji Federated States of Micronesia Papua New Guinea Samoa Solomon Islands Vanuatu Other countries Brunei Darussalam Cambodia Malaysia Mongolia Philippines Singapore Viet Nam TOTAL All the subgroups of countries reduced their regional collaboration as the visibility of the publications increased, and two countries seemed to have escaped this reality. The first is Australia, which maintained its share of global production for publications with low visibility (2.35%) and increased its share for publications with high visibility (2.49%). In fact, its share of publications with very high visibility (2.06%) was relatively close to its share in global output. The other is Papua New Guinea, which also increased its global share in high visibility publications (0.09%). The former, however, seems to have had more statistical robustness because of the relatively large number of publications that were produced. The other country that was relatively close to maintaining its share in high visibility publications is New Zealand. Yet no other country comes close to this, including other important producers such as China, Japan and the Republic of Korea. In all cases, their research production tended to be concentrated in the low visibility category. 11 If a region was balanced with regard to the patterns of visibility, its production for each level of visibility would reflect the same structure of its total production, and its indexes of visibility would be equal to 1 in all cases. 12 If one analyses the distribution of publications by visibility level and income group (see Table A13 in the Statistical Annex), the countries of low and medium (medium-high and medium-low) incomes concentrate their production in publications with low visibility. There is a rather high correspondence between income group and regions as well. 13 To construct this table, information was taken from Table A13 in the Statistical Annex 19

27 4.2 Visibility and collaboration in the Western Pacific Region Why did some countries drastically reduce their participation in publications as their visibility increased? Although the information contained in the analysed database fails to provide a definitive answer, some hypotheses can be advanced. Perhaps publications that originated in certain countries, e.g. Australia, were of greater quality or were published in more visible journals that have better quality standards. The second hypothesis is that, due to specific research-oriented policies, the scientific networks were much denser in countries like Australia or New Zealand. Thus, members of these networks not only could collaborate among themselves with greater ease but, in addition, they cited their own works more frequently. This, ultimately, led to an increase in visibility. If this occurred, a possible strategy for researchers from other countries is to integrate these networks. Collaboration with high visibility countries would be the fastest (and less expensive) way to achieve higher visibility. 14 Table 17 shows collaboration patterns of countries in the Western Pacific Region according to publications visibility for 1992 and The table shows the effect that different types of collaborations may have had on the visibility of the research output. The first line ( No collaboration ) refers to publications achieved within a single institution. National collaboration refers to publications that engaged only two or more institutions within the same country. International collaboration refers to publications that had at least two institutions from different countries. This category is broken down into collaboration exclusively between countries in the Western Pacific Region and between at least a Western Pacific Region country and at least an extraregional country. In turn, this category is further split into collaborations with at least one extraregional high-income country and with other non-high-income countries. Table 17. Collaboration and visibility in the Western Pacific Region Visibility of research: below field average Publications Percentage Publications Percentage No collaboration National collaboration International collaboration Intraregional only With at least one extraregional country With at least one high-income country With no high-income country TOTAL Visibility of research: above field average Publications Percentage Publications Percentage No collaboration National collaboration International collaboration Intraregional only With at least one extraregional country With at least one high-income country With no high-income country TOTAL Naturally, it is not simple to be integrated in these networks. Usually, the invitation to become part of them comes from the members of such networks (scientists in developed countries) and, naturally, the research topics and agendas are determined by the interests of these members.

28 From Table 17, publications with no collaboration decreased between 1992 and 1998, both for low (below field average) and high visibility publications (above field average). Both national and international collaboration increased during this period for both types of visibility. This pattern is entirely consistent with Section 3. Secondly, if what researchers sought was visibility for their research, projects without scientific collaboration were not a good means to achieve it. During 1998, almost 45% of low visibility publications were produced without collaboration, but only 31% of those with high visibility were written this way. Hence, individual efforts (at the author or institution level) did not seem to take full advantage of these networks by which publications could increase their visibility. Publishing alone did not help to show the publications. Thirdly, and according to the information contained in Table 17, there were two ways to increase visibility of research output. The first and most effective was to collaborate with an extraregional country. While during 1998, 14% of low visibility publications were achieved with extraregional countries collaboration, such a figure increases to 29% for high visibility publications. Within this type of collaboration, the most successful way of achieving high visibility was to collaborate with a high-income country. Finally, Tables 18 and 19 display the relationship between collaboration and visibility for the five largest producers, and all of them, except China, were high-income countries. Table Visibility of research, below field average 1992 Australia China Japan Rep. of Korea New Zealand No collaboration or national collaboration International collaboration Within Western Pacific Region only Outside Western Pacific Region exclusively Only with high-income countries With at least some high-income country Within and outside Western Pacific Region TOTAL (Percentage) Total number of publications Australia China Japan Rep. of Korea New Zealand No collaboration or national collaboration International collaboration Within Western Pacific Region only Outside Western Pacific Region exclusively Only with high-income countries With at least some high-income country Within and outside Western Pacific Region TOTAL (Percentage) Total number of publications Comparing both tables results, the percentage of publications with no collaboration or with national collaboration was far higher for low visibility research (Table 18) than for high visibility research (Table 19). It is clear from both tables (as it is clear for other regions) 16 that collaboration with other countries increased the likelihood of having more citations, possibly because this type of collaboration is the easiest and more direct way of accessing by several scientific networks. 15 Taken from Paraje, S. and Salmela R. Collaboration and visibility of health research in the Western Pacific Region, presented at the Global Forum 11, Beijing, China, Paraje, G. Production of scientific articles in the Region of the Americas, presented at the XXXIX Pan American Health Organization Advisory Committee of Health Research, Santiago, Chile, November

29 Table Visibility of research, above field average 1992 Australia China Japan Rep. of Korea New Zealand No collaboration or national collaboration International collaboration Within Western Pacific Region only Outside Western Pacific Region exclusively Only with high-income countries With at least some high-income country Within and outside Western Pacific Region TOTAL (Percentage) Total number of publications Australia China Japan Rep. of Korea New Zealand No collaboration or national collaboration International collaboration Within Western Pacific Region only Outside Western Pacific Region exclusively Only with high-income countries With at least some high-income country Within and outside Western Pacific Region TOTAL (Percentage) Total number of publications Visibility tended to be higher for collaboration with countries outside the Region. In almost all cases, the proportion of high visibility research conducted in collaboration with countries outside the Western Pacific Region at least doubled low visibility research. By 1998, for all countries (except Japan), around one in three high visibility papers was written with extraregional collaboration. From 1992 to 1998, this proportion rose for all countries except China and the Republic of Korea. Finally, a large proportion of high visibility extraregional collaboration occurred with high-income countries. Between 72% (New Zealand) and 87% (Japan) high visibility production fell into this category during Though collaboration with non-high-income countries slowly rose (except for Republic of Korea, where it fell between 1992 and 1998), the bulk of this research was done with wealthy countries, frequently without the involvement of other countries. For 1998, high visibility research publications between 69% (New Zealand) and 84% (Japan) were written in collaboration with extraregional high-income partners. 17 Taken from Paraje, S. and Salmela R. Collaboration and visibility of health research in the Western Pacific Region, presented at the Global Forum 11, Beijing, China,

30 5. PUBLIC HEALTH RESEARCH PRODUCTION IN THE WESTERN PACIFIC REGION This section describes and analyses the patterns of research production and regional collaboration for publications that appeared in scientific journals classified in the public health and health systems (PHHS) category (see Table 1). Despite the statistical noise from imperfections in Thomson ISI s classification system, observation of the main trends show interesting conclusions. 18 Table 20 shows the regional distribution of global publications in PHHS. As can be seen from the table, the main contributors in this area were the Region of the Americas, the European Region and the Western Pacific Region. The Western Pacific Region s participation was significantly lower than its participation in total health research production (around 13%, as shown in Table 2) marking the relatively low importance that this research area had in the Region. Table 20. Research production in public health and health systems by WHO region Average WHO Regions Region of the Americas African Region Eastern Mediterranean Region European Region South-East Asia Region Western Pacific Region The South-East Asia Region and the African Region shared the fourth position, with 2% averages. The case of the African Region is revealing since its participation in PHHS literature almost tripled its participation in health in general. It is probable that this Region s enormous public health problems make institutions in developed countries seek collaboration with African Region countries in order to have a local platform of study. Public health research has certain special features that favour developing countries, e.g. their knowledge of local realities and field studies that often require the participation of local institutions. Table 21 shows the participation of each group of countries, by income level, in PHHS publications. Low-income countries had an important participation rate of 3.3% of the total PHHS output. This percentage was sufficiently higher than that registered for health in general, which was close to 1.7%. 19 This also reflects the relative importance of this group of countries when PHHS is the principal research topic. Table 21. Research production in public health and health systems by income group in the Western Pacific Region Average Low income countries Low-middle income countries High-middle income countries High income countries Within the Western Pacific Region, however, the distribution PHHS output was very different from that for health in general (see Table 6). Table 22 shows that Japan was the main PHHS producer with only one third of the regional production. Though this share is not unimportant, it was lower than its share in regional health- related production. Perhaps the low share in Japan s contribution to PHHS was related to its relative closeness. As demonstrated in Section 3, Japan did not collaborate 18 There is an additional factor that should be taken into account when evaluating this section s results: the number of journals included in such a category and their geographical representation. As the analysis is restricted to specific scientific areas, the number of scientific journals analysed is reduced and, therefore, any small change in the composition of this sample can affect the results. For example, the inclusion or exclusion of a journal can affect the estimates obtained on scientific output for a specific region or country. In this respect, results should be treated with extreme caution. 19 See Paraje G., Sadana R. and Karam G. Increasing international gaps in health-related publications, 2005, Science 308: The gain registered by low-income countries in terms of their share in PHHS scientific production is compensated by the loss of high-income countries. 23

31 much with other regional countries or low-income extraregional countries. As was suggested, if PHHS research is frequently linked with problems arising in least developed areas, not collaborating with them can prevent such research from occurring. Other countries, such as Australia, China and New Zealand, had relatively large shares in PHHS research. Australia contributed with a third of the regional production, whereas China and New Zealand contributed with 19% and 6%, respectively. In the cases of Australia and New Zealand, this participation doubled their regional participation in total health-related research, showing how strong they were in this particular scientific area. Not surprisingly, both Australia and New Zealand were two of the most open countries to collaboration within the Region and with relatively poorer countries. Table 22. Research production in the public health and health systems category in the Western Pacific Region, percentage of regional production Average Countries Five largest producers Australia China Japan Republic of Korea New Zealand Pacific Islands Cook Islands Fiji Marshall Islands Federated States of Micronesia Palau Papua New Guinea Samoa Solomon Islands Tonga Tuvalu Vanuatu Other countries Brunei Darussalam Cambodia Lao People s Democratic Republic Malaysia Mongolia Philippines Singapore Viet Nam TOTAL On the contrary, some other countries reduced their participation in PHHS. The most notable, aside from Japan, is the Republic of Korea, which decreased its contribution to less than half (7.1% in case of health in general, to 3.1% in PHHS). Pacific islands increased their regional share more than 10 times although their regional contribution was still well below 1.0% (on average around 0.6%). Other regional producers were doubling their participation from 2.0% for general health-related research to 4.1% for PHHS research. The main contributor was Singapore, with 1.8% of regional production. However, Malaysia had a strong research programme in this area as well, as its share in PHHS research more than doubled its share in general health-related research (from 0.4% to 1.1%). The Philippines and Viet Nam also showed important increases. 24

32 Finally, Table 23 shows the patterns of PHHS collaboration of countries in the Region with the rest of the world. The table, which should be compared to Table 11, shows specific aspects that must be emphasized. Table 23. Collaboration patterns for PHHS research production Collaboration type Publications Percentage Publications Percentage Intraregional collaboration (1) Extraregional collaboration (2) with high income countries with high-middle income countries with low-middle income countries with low income countries No collaboration (3) TOTAL (1) Collaboration between at least two institutions from the Western Pacific Region only. (2) Collaboration with at least one institution from other regions. (3) Only a single institution involved in the publication. In both PHHS and general health-related research between 1992 and 2001, noncollaborative research production suffered a steep decline. In 1992, PHHS research represented 52% of research production; in 2001, it represented 33%. This finding is compatible with Table 11. Also consistent with Table 11 are the patterns followed by intraregional and extraregional collaboration on PHHS. The first type of collaboration increased between 1992 and 2001 (more than 10%), whereas the second increased 9% in the same period. In the case of general health-related research production, the increase in extraregional collaboration was not as significant. Collaboration with high-income countries was the main factor behind such an increase. Unlike the case of general health-related research, in PHHS research there was a relative decrease in collaboration with lowincome countries. 25

33 STATISTICAL ANNEX Table A1. Health-related research production in the Western Pacific Region, percentage of world production Countries Average Five largest producers Australia China Japan Republic of Korea New Zealand Small Pacific Islands Cook Islands Fiji Marshall Islands Federated States of Micronesia Palau Papua New Guinea Samoa Solomon Islands Tonga Tuvalu Vanuatu Other WPRO countries Brunei Darussalam Cambodia Lao People s Democratic Republic Malaysia Mongolia Philippines Singapore Viet Nam TOTAL Table A2. Health-related research production in the Western Pacific Region, percentage of world production Countries Average High income countries Australia Brunei Darussalam Japan Republic of Korea New Zealand Singapore High-middle income countries Malaysia Palau Low-middle income countries China Cook Islands Fiji Marshall Islands Federated States of Micronesia Philippines Samoa Tonga Tuvalu Vanuatu Low income countries Cambodia Lao People s Democratic Republic Mongolia Papua New Guinea Solomon Islands Viet Nam TOTAL

34 Table A3. Australia: Patterns of research collaboration, 2001 Countries with collaborations Publications Percentages United States of America United Kingdom Germany Canada France Japan New Zealand China The Netherlands Italy Rest of the world Number of publications with international collaboration Table A4. China: Patterns of research collaboration, 2001 Countries with collaborations Publications Percentages United States of America Japan United Kingdom Canada Germany Australia France Sweden The Netherlands Singapore Rest of the world Number of publications with international collaboration Table A5. Japan: Patterns of research collaboration, 2001 Countries with collaborations Publications Percentages United States of America United Kingdom Germany China Canada France Republic of Korea Australia Sweden Italy Rest of the world Number of publications with international collaboration

35 Table A6. Republic of Korea: Patterns of research collaboration, 2001 Countries with collaborations Publications Percentages United States of America Japan Canada United Kingdom China Germany Sweden Italy France Australia Rest of the world Number of publications with international collaboration Table A7. Malaysia: Patterns of research collaboration, 2001 Countries with collaborations Publications Percentages United Kingdom United States of America Japan Australia Indonesia Singapore India China France New Zealand Rest of the world Number of publications with international collaboration 152 Table A8. New Zealand: Patterns of research collaboration, 2001 Countries with collaborations Publications Percentages United States of America Australia United Kingdom Germany Canada Japan Sweden China France Italy Rest of the world Number of publications with international collaboration

36 Table A9. Papua New Guinea: Patterns of research collaboration, 2001 Countries with collaborations Publications Percentages Australia United States of America Switzerland France Germany Italy United Kingdom Rest of the world Number of publications with international collaboration 48 Table A10. Philippines: Patterns of research collaboration, 2001 Countries with collaborations Publications Percentages United States of America Japan Germany Australia China India Thailand Republic of Korea Canada Indonesia Rest of the world Number of publications with international collaboration 123 Table A11. Singapore: Patterns of research collaboration, 2001 Countries with collaborations Publications Percentages United States of America United Kingdom China Australia Canada Japan Germany Switzerland France India Rest of the world Number of publications with international collaboration

37 Table A12. Viet Nam: Patterns of research collaboration, 2001 Countries with collaborations Publications Percentages United States United Kingdom France Japan Sweden Australia Germany Thailand Belgium China Rest of the world Number of publications with international collaboration 169 Table A13. Research production by visibility, Western Pacific Region countries grouped by income, 1998 Subgroups / Countries Share in global research output (%) Lower than 1 Ratio of expected citations Between 1 and 2 Greater than 2 High income countries Australia Brunei Darussalam Japan Republic of Korea New Zealand Singapore High-middle income countries Malaysia Low-middle income countries China Fiji Federated States of Micronesia Philippines Samoa Vanuatu Low income countries Cambodia Mongolia Papua New Guinea Solomon Islands Viet Nam TOTAL

38 Table A14. Research production by visibility, regional shares of countries in the Western Pacific Region Subgroups / Countries Share in global research output (%) Lower than 1 Ratio of expected citations Between 1 and 2 Greater than 2 Five largest producers Australia China Japan New Zealand Republic of Korea Small Pacific Islands Fiji Federated States of Micronesia Papua New Guinea Samoa Solomon Islands Vanuatu Other WPRO countries Brunei Darussalam Cambodia Malaysia Mongolia Philippines Singapore Viet Nam TOTAL Table A15. Public health and health systems research production in the Western Pacific Region, percentage of global production Subgroups / Countries Average Five largest producers Australia China Japan New Zealand Republic of Korea Small Pacific Islands Cook Islands Fiji Marshall Islands Federated States of Micronesia Palau Papua New Guinea Samoa Solomon Islands Tonga Tuvalu Vanuatu Other WPRO countries Brunei Darussalam Cambodia Lao People s Democratic Republic Malaysia Mongolia Philippines Singapore Viet Nam TOTAL

39

40