Research and Collaboration Overview of Institut Pasteur International Network: A Bibliometric Approach Toward Research Funding Decisions

Background: Institut Pasteur International Network (IPIN), which includes 32 research institutes around the world, is a network of research and expertise to fight against infectious diseases. A scientometric approach was applied to describe research and collaboration activities of IPIN. Methods: Publications were identified using a manual search of IPIN member addresses in Science Citation Index Expanded (SCIE) between 2006 and 2011. Total publications were then subcategorized by geographic regions. Several scientometric indicators and the H-index were employed to estimate the scientific production of each IPIN member. Subject and geographical overlay maps were also applied to visualize the network activities of the IPIN members. Results: A total number of 12667 publications originated from IPIN members. Each author produced an average number of 2.18 papers and each publication received an average of 13.40 citations. European Pasteur Institutes had the largest amount of publications, authored papers, and H-index values. Biochemistry and molecular biology, microbiology, immunology and infectious diseases were the most important research topics, respectively. Geographic mapping of IPIN publications showed wide international collaboration among IPIN members around the world. Conclusion : IPIN has strong ties with national and international authorities and organizations to investigate the current and future health issues. It is recommended to use scientometric and collaboration indicators as measures of research performance in IPIN future policies and investment decisions.

approach. These findings can assist IPIN administrators and stakeholders in allocating funds, benchmarking and accomplishing their missions.

Scientometric analysis
This study was designed based on Science Citation Index Expanded (SCIE) members' scientific publications in scholarly journals. An SCIE search was performed in June 2012. Filters were used to include IPIN members' relevant publications using organization enhanced panel. Articles published between 2006 and 2011 were obtained by scanning their address field (with at least one author affiliated to an IPIN member), under the heading "Pasteur". This review resulted in publications originating from the IPIN members, and then each article was subcategorized by geographic regions (based on World Health Organization (WHO) regions). Absolute Number of Articles (ANA), Total Number of Citations (TNC), Citation to any Papers (CP), Average number of Authors per Papers (AAP), Average number of Papers to Author (APA), and the H-index were used to estimate the scientific production of each IPIN member. A comparison of IPIN members was presented based on the above-mentioned indicators. IPIN includes institutions from five continents. However, in this study, all IPIN members were sub-classified into geographical regions based on WHO classification to increase the efficacy of IPIN ranking and to better describe the position of each Institute in the region.

International collaboration
The publications of IPIN were separately retrieved according to the name of the institute and research areas. Data were analyzed using ISI.exe (20,21) in the first step and Cities1.exe (21) afterwards. The retrieved addresses of the Pasteur Institutes and other collaborating organizations were extracted using GPS visualizer (22) and as a result, the precise addresses of the cities and institutes were identified (21). Using Pajek software (23), the output of this file was then demonstrated in the form of a figure representing special fields with the most frequently studied topics among IPIN members. In addition, Google Maps and its facilities were applied to reflect the geographic mapping of the collaborative scientific networks according to their dominant fields of activity. VOSviewer was also used to visualize the network map of IPIN international collaboration (24).

Webometric analysis
In order to compare the webometric indicators of the network members in 2012, the activity and impact indicators of the web were applied (25). To reach this goal, the status of the Pasteur Institutes websites were assessed via link analysis according to the standard criteria of webometrics, Impact (backlinks and referring domains), Size of website and the number of rich files (the number of documents which are indexed with the website address of the Pasteur Institutes in Google Scholar and SCImago SIR) (26). In this study, the Pasteur Institutes were compared according to the activity, impact indicator, and the main webometric indicators (a combination of activity and impact indicators). According to the methodology of the world webometric ranking, the composite indicator was built for any institute based on impact (50%) and activity (50%). Table 1 provides more details on the weighting criteria and composites of webometric indicators.

Statistical Tests
All tests were conducted using SPSS 16 (SPSS Inc., Chicago, IL, USA). Pearson correlation analysis and one-way analysis of variance (ANOVA) test were used for further analysis. P-value less than 0.05 was considered statistically significant. International collaboration IPIN members have collaborated with 726 non-pasteurian institutes around the world. The European IPIN members had the largest amount of collaboration among IPIN members ( Figure 2 and Figure 3). The overall collaboration followed a similar pattern in main subject areas. Biochemistry and immunology collaboration patterns were more similar in international connections among European and American institutions; however, collaboration among Asian and European countries were more significant in microbiology. Infectious diseases revealed broader and wider connections across the globe rather than others. It seemed that European institutions received more connections and were major parts of international collaboration among different subject categories. However, these connections were more distinct between American and European institutions. Figure 3 indicates the overall collaboration of IPIN members according to their number of co-authorship links.

Webometric analysis
All IPIN members had an independent website, except Laos. Amongst the rest of the Pasteur Institutes, the highest impact indicators belonged to the Pasteur Institutes of Paris, Lille, and Iran, while the highest activity indicators belonged to the Pasteur Institutes of Paris, South Korea and Belgium. The highest ranking according to the total score of webometric indicators was given to the Institut Pasteur of Paris, Lille and Iran (Table 2). Pearson correlation analysis showed a positive significant relationship among the Pasteur Institutes around the world regarding activity and impact indicators (P= 0.05, r= 0.44). Moreover, Pearson correlation analysis also showed a positive significant relationship between the total score of webometrics and the indicators including the total number of articles (P= 0.01, r= 0.57), the number of citations (P= 0.01, r= 0.55), and the H-Index (P= 0.01, r= 0.63). Comparing the webometric indicators among different geographic areas revealed that the European Pasteur Institutes were the pioneer in their activities and impact indicators in websites, followed by the Asian Pasteur Institutes. The lowest level of these indicators belonged to the African Pasteur Institutes (Table 3). However, no statistically significant differences was observed among geographical regions.

Discussion
This study demonstrated the unique role of IPIN European members in research and network activities according to scientometric and webometric indicators. Major themes of this work is represented and highlighted with a conclusion in the following sections.

IPIN's publication patterns
The findings of this study indicated that the European members especially Paris, Lille, and Greece were the most considerable institutes which have been demonstrated to be an important production source of scientific publication, because of their developments in basic and applied research (4). African members obtained lower ranks in IPIN research activities that might be due to their shortcomings, language and the lower quantity and quality of their articles (27,28). In terms of rank, the European Pasteur Institutes were placed in the first rank, followed by the Pasteur Institutes of Vietnam, Iran, Tunisia and South Korea. Although in this study the highest scientometric indicators belonged to the European members, the average number of these indicators was not significantly different between European and non-European Institutes. This fact can be simply explained by the lower rank of the Pasteur Institutes in Eastern Europe (Bulgaria, Russia, and Romania) compared to those in the central and Western Europe (France (Lille), Greece, Italy, and Belgium). Interestingly, there were notable similarities between the findings of the previous studies and our findings considering the number of affiliated authors as a major determinant of the scientific output indicators regardless of history (29). Eastern Mediterranean region is one of the most leading sites of IPIN in terms of research activities, placed after European members in higher ranks of studied indicators. It seems that Iran, Tunis, and Morocco have a remarkable place among IPIN non-European members. These institutions are considered as the old members of the network. Among these, Iran is the only institute located in Asia according to the Iranian Ministry of Health's annual reports. Pasteur Institute of Iran is one of the most productive research institutes in the country. It is the only research institution in the country which is ranked among universities (30). With that in mind, Pasteur Institute of Iran is one of the most influential and productive institutions in national and regional perspectives. Institut Pasteur of Paris was the most considerable member in IPIN for its international achievements and scientometric indicators possibly due to its larger number of authors.

IPIN's international collaboration
Geographical mapping of IPIN scientific collaboration in major research areas signified a rather similar trend of collaboration among the members. Despite the differences in IPIN research and network activities, collaboration were typically formed between the European members based on their relative   geographical proximity. This could be due to many factors. For example, researchers may be more interested in forming scientific collaboration with their colleagues from nearby countries, because they may encounter each other at regional conferences more often. Besides it may be easier to collaborate with researchers sharing the same language or culture (31). European Union's policies in improving the scientific collaboration among European members may be another reason (32). It seems that Europe, northern America, South-east Asia and western Pacific are to some extent the most collaborative regions in the world. All Pasteur institutes around the world are more likely to collaborate with European members specially Institut Pasteur of Paris as a core element in the network, forming a centralized network. Preliminary studies have indicated several reasons for international collaboration among different countries. In this regard, sharing knowledge and transmitting information, access to resources and equipment, higher quality of the research, and sharing costs were identified as motivating factors (33,34). However, developments and the economic status of countries could make different patterns of collaboration across the world. Evidence showed that international collaboration of peripheral and developing countries is directed to attain knowledge and techniques from developed countries (33,35). Getting more citations and better prestige were mentioned as other encouraging reasons (33). Our findings revealed that except for infectious diseases, other subject domains indicated major connections between European, American, and Eastern Asian countries. Since countries located in South-east Asia are among developing and low-income countries, it could be understood that their connections were to get more knowledge and advanced equipment, as other developing countries. It was also noted that IPIN members revealed few connections within Thus, it is highly suggested that IPIN regional offices, play an important role in improving regional collaboration among nearby members, sustaining domestic capabilities and research expertise in the regions to foster regional reactions in the case of emergent conditions.

IPIN's citation patterns
Increasing scientific collaboration would lead to an increase in the scientific production and quality of publications. Therefore, citations to those articles would go up (32,36). That is why European institutions had the highest scientometric indicators besides the highest level of international collaboration. As shown in Figure 1, the European Pasteur Institutes had significant collaboration with the developing IPIN members (particularly in Africa and Western Pacific). It could be simply explained by the need of African and Asian countries to the potential facilities and capabilities of the developed countries, and the policy of European countries in developing international collaboration. As the scientific and international collaboration of the European Union members with non-European countries was significantly increased in 2005 compared to 1997-1999 (37,38), the interest of some Asian countries to have more collaboration with Europeans have been mentioned by recent works. According to the National Science Foundation of America (NSF)'s report, international collaboration of Asian countries with non-Asian ones has annually increased by 11% over recent years, showing their inclination toward scientific collaboration with developed countries, particularly European ones (27). Increasing the scientific collaboration among countries is a progressive trend in science. Lack of expert human forces and resources, against the increasing need of conducting multidisciplinary and interdisciplinary research are the most important reasons (32,37,39). Since IPIN members are united by the same Pasteurian culture, the same scientific rigor and the same values, it is expected to have closer scientific relationship with each other within the network. But this finding signified that IPIN members prefer to have close collaboration with European members, especially Paris, rather than having a wide range of connection through all IPIN members. Improving the scientific collaboration among IPIN members as well as non-Pasteurian institutes will improve the IPIN publications' impact. It also increases the quality of researches they conduct. More collaborative members of IPIN will receive much attention and impact from the scholar community, if IPIN administrations have a look on network   IPIN is one of the most known research networks working on infectious diseases by producing vaccines, developing immunization programs, and having a broad impact on the public health promotion worldwide (12).

Conclusion
Having global collaboration, European Pasteur Institutes especially Institut Pasteur of Paris, Lille, and Greece were the most noticeable IPIN members. Eastern Mediterranean members including Pasteur Institutes of Iran, Tunis, and Morocco were after the European members. These members were considered as significant ones in the IPIN organization. Comparing Pasteur Institutes according to their highly valued parameters help the IPIN administrations modify their policies based on the best available evidence provided by scientometric indicators. Our findings recommended using scientometric and collaboration indicators as measures of research performance in IPIN future policies and investment decisions. Furthermore, establishing sub-regional networks and improving regional offices might serve better in the quality of regional collaboration among IPIN members and fulfilling the regional requirements as well.

Ethical issues
This study was funded by the Pasteur Institute of Iran as a research grant.

Implications for policy makers • Institut Pasteur International Network (IPIN)
collaboration patterns are mostly presented as tight connections among American, European and Eastern Asian Institutions. • Research and collaboration indicators of IPIN members are proposed as tools in evaluating research performance and allocating research funds within the network. • Presentation of collaboration patterns among research collaborative members is informative to explore how these members work together towards their common goals. • A world snapshot on IPIN collaboration patterns in major research domains is provided to indicate how IPIN members work together towards their common goals.
Implications for public Institut Pasteur International Network (IPIN), since 1888, is intended to fight against infectious diseases. Now with 32 members around the world, IPIN provides a unique capacity for research and public health network collaborations to the prevention and treatment of infectious diseases and promotion of public health. IPIN members are united and have close scientific relationship with each other within the network. They share the same Pasteurian culture, the same scientific rigor and the same values. Biochemistry and molecular biology, microbiology, infectious diseases and immunology are the IPIN's major research domains. IPIN members prefer to have close collaboration with European members, especially Paris; they also have collaboration with 726 non-Pasteurian institutes around the world. Such examples of successful cross-national research collaborations would herald an immense boost in public health promotion worldwide.