Virology research in a Latin American developing country : a bibliometric analysis of virology in Colombia ( 2000 – 2013 )

Introduction: Bibliometric analysis demonstrates that the virology research in Latin America has increased. For this reason, the objective of this study was to evaluate Colombian publications on viruses and viral diseases in indexed journals during the period from 2000 to 2013. Methodology: The bibliographic data were collected from MedLine, SciELO, LILACS and Scopus databases. The database was constructed in Excel descriptive statistics. The SCImago Journal Rank (SJR) was evaluated using the SCImago Journal & Country Rank in 2013 and was used as an indicator of the quality of the journals used by the Colombian researchers. Results: The total number of papers published was 711, of which 40.4% were published in local journals, and 59.6% were published in foreign journals. Most (89.2%) were original papers. Moreover, 34.2% of the papers were published in collaboration with international researchers, with the United States being the most represented. Of the journals used, 85.6% had an SJR, and 14.4% did not. The median SJR of the papers was 0.789, and the median of the papers with international collaborators was higher compared to the SJR of the papers without international collaboration. Papers were most frequently published in journals whose categories were medicine (miscellaneous), virology, and infectious diseases. The viruses that appeared in the papers more frequently were HIV, dengue, and papillomavirus. Conclusions: This study provides data for use in research, health planning, and policy analysis as it relates to virology in Colombia and other developing Latin American countries.


Introduction
Virology is a field of microbiology that studies a broad group of viral diseases that are the cause of considerable morbidity and mortality worldwide in different human, animal, and plant populations.Virology research ranges from the clinical findings, the etiology, pathogenesis, epidemiology, prevention and treatment of viral diseases, to the molecular therapeutic for cancer and other viral and non-viral diseases [1].
According to International Committee on Taxonomy of Viruses (ICTV), viruses comprise seven orders, 87 families, 19 subfamilies, 349 genera, and 2,284 viruses and viroid species [2,3], and are becoming the most diverse microorganisms on earth.
Several investigators have conducted bibliometric analyses of research productivity of different regions of the world, focused mainly on biomedical fields [4,5,6].These bibliometric analyses have shown that research in microbiology [7] and research in virology [8] are concentrated in developed areas (United States and Western Europe), which have produced the majority of the world's virology research in terms of both quantity and quality of information.Specifically, these two world regions have produced 77.7% of the published articles in this field [8].
From previous bibliometric analyses, is interesting to note that between the years 1995 and 2003, virology research in Latin America exhibited a tendency to a relative increase of production in published articles indexed in the Journal Citation Reports database of the Institute for Scientific Information (ISI) [8].Similarly, a recent analysis in Colombia evaluating publications on infectious disease publications in Colombian journals showed a similar increase between 2000 and 1227 2009 and noted that there is an imbalance between different areas of the country [9].
According to previous analysis, there is an urgent need for virology researchers from developed regions to enhance their research cooperation with developing areas to perform the most accurate science and to have more equity in these developing regions [8].However, it is important to highlight that Falagas et al.only evaluated articles belonging to the virology category [8], and it is clear that due to wide importance of virology, this approach creates a gap in the knowledge, because many articles regarding viral diseases are published in journals of other categories such as medicine, general and internal; medicine, research and experimental; infectious diseases; veterinary (miscellaneous); agronomy and crop science, and others.
It is important to highlight that cooperation between countries in Latin America has been proposed as an alternative way to obtain equity in these developing regions.Regional initiatives, such as the Brazilian/Argentinean Center for Biotechnology (CABBIO), have increased the cooperation and the publication index between those countries and others from the region through the implementation of training courses in biotechnology and grant funding mainly in the Mercosur area [10,11].The CYTED initiative (from the Spanish Programa Iberoamericano de Cooperación en Ciencia y Tecnología para el Desarrollo: Ibero-American Cooperation for Scientific and Technological Development Programme) founded by Latin American and Caribbean countries, Portugal, and Spain to enable further cooperation among research terms, has led to important advances in increasing scientific cooperation and development [12,13].However, only one initiative in virology has been placed in the virology area in all Ibero-America, the Virored-Cyted network (Red Temática en Virosis Emergentes: Thematic Network on Emerging Viruses), which was created in 2010 and, unfortunately, has been inactive since 2013 [14].
In recent years, no one has quantified the real balance of virology-related research by Colombian researchers in indexed journals.It is necessary to evaluate if scientific production in virology has continued to increase during the last decade and if the indicators show a real potential for Colombia in basic science and clinical aspects of viral diseases.This paper evaluates Colombian publications on viruses and viral diseases in indexed journals covering several areas of virology and published in many different journal categories to provide basic data for use in research, health planning, and policy analysis regarding virology in Colombia and other developing Latin American countries, as has been proposed in other Latin American bibliometric initiatives [15].

Data collection
This retrospective and documental study was conducted using printed and Epub ahead-of-print online articles published from January 2000 to December 2013.The bibliographic data were collected from the databases of the National Library of Medicine on the web (MedLine) (http://www.ncbi.nlm.nih.gov/pubmed), the Scientific Electronic Library Online (SciELO) (http://www.scielo.org),LILACS, the Pan-American Health Organization's supported database on Latin American and the Caribbean Literature in Health Sciences (http://lilacs.bvsalud.org/en/),and Scopus database from Elsevier (http://www.scopus.com/).The data were exported to Excel, and the database was constructed with the title, author, address, source, year, type of publication, journal title, language of publication, virus, journal category, and SCImago Journal Rank (SJR) of the journal.The SJRs were obtained from SCImago Journal & Country Rank and were arbitrarily adopted to estimate the visibility of the papers and the quality of the journals [15,16].The SJR was used as an indicator of the quality of the journals used by the Colombian researchers.As reported by others, when using the descriptive statistics of SJR for research documents, those with no SJR were given a score of zero [17].To clarify international cooperation, the collaborating country was included in the database search, and the absolute country counting method was adopted [18].The citation analyses were retrieved only from the Scopus database and obtained as a table with numbers of cited articles for individual years, as well as the total number of cited references for all years [19].Notes, letters, editorials, news, and meeting abstracts were excluded.

Search strategy
For each search in the PubMed database, a sentence consisting of different parts joined together by Boolean operators, i.e., AND, OR, and NOT was used in the search field.By making use of the Limits function that is incorporated in the search engines, each search was limited to specific years.Manual detection was used to exclude publication types, such as notes, letters, editorials, news, and meeting abstracts.The words Colombia AND virus were selected as medical subject heading (MeSH) terms.Despite this meticulous search, some articles were missed because the full address was not registered.In order to include these articles, the words that could help to find several important viral diseases that specifically affect Colombia were included, such as dengue, HIV, Solanum  The search in Scopus was performed using the basic search system and the date range to delimitate the search.A mix of words joined together by Boolean operators were used in the basic search system (article title, abstract, key words).An example of the terms used in the main search follows: virus OR viruses OR dengue OR HIV OR rabies OR HPV AND Colombia.
The SciELO and LILACS databases were accessed through the databases' websites and the review was operationalized through electronic search of articles indexed in the subject index [20] based on the key words Colombia and virus, and using date limits.Finally, the Excel database was filtered to eliminate repeated records from different primary databases.To strengthen the methodological validity of this study, two independent authors conducted the data collection and filtered the results, as previously reported [21].In case of disagreement between the two investigators, the results were discussed in research group meetings.

Statistical analyses
Descriptive statistics were used in all cases to evaluate the results (i.e., the frequency in count and percentage).To measure the tendency in total number of published papers by year, lineal regression (R 2 ) was used as a measure of lineal association with a p value of < 0.05.The Shapiro-Wilk test was used to evaluate normality, and differences in SJR between papers with and without international cooperation were evaluated using the Mann-Whitney test.A p value < 0.05 was regarded as statistically significant (two-tailed).Statistical analyses were performed using GraphPad Prism version 6 for Windows.

Results
There were a total of 711 papers published by Colombian researchers in different areas of virology that were included in the present study.Figure 1 presents the actual number of articles produced each year during the study period.It is clear that the total number of papers increased progressively during the period (r 2 = 0.9467; p < 0.001).There was a more than eightfold increase in the number of papers from 2000 (12) to 2013 (103), with the highest number of published papers in 2013 (Figure 1A).
Taking into account that databases included local indexed journals (Colombian journals), the number of papers published in foreign (international) or national (local) journals was evaluated.During almost all the years studied, papers in foreign international journals accounted for the majority of publications, ranging from 37% (2004) to 73% (2001) with an average of 59.6% of international publications from 2000 to 2013 (Figure 1B).There was no statistically significant correlation between the number of articles appearing each year and the percentage of articles published in international journals.
Because only primary research, which includes mainly original papers, short communications, short papers and reviews/assays was included, the percentage of original research of Colombian researchers and the total number of reviews or essays as possible indicators of top research in different areas of virology was sought.Of the published papers, 89.2% were original papers and only 10.8% percent were reviews and/or essays.
An important finding was that the publication language followed the same pattern as the international publications, with 52.9% of the papers written in English, 46.8% in Spanish, and only 0.3% written in Portuguese.No paper appeared in another language.
Furthermore, 94.9% of the publications were coauthored.The median number of authors was 5.3 (range, 1 to 38).On average, articles with 10 or more authors appeared in journals with higher SJRs than did those with fewer than four authors.Table 1 presents the total distribution of the number of authors per publication.
There is an urgent need for Latin American researchers to enhance their cooperation with developed countries to enhance their research capacity.With this in mind, foreign collaboration with Colombian researchers was evaluated.A total of 34.2% (243) of the papers published during the study period included foreign international collaboration, as indicated by the authors in the papers (rank between 20.0% in 2004 and 41.7% in 2000); in some of them, the researchers collaborated with more than one country or international institution.The tendency can be observed in Figure 2A.
The main collaborating countries were the United States, Spain, and France (42.4%, 13.6%, and 10.7%, respectively), which account for more than a half of the studies with international cooperation.However, Colombian researchers collaborated with 45 different countries, including some important Latin American productivity leaders such as Brazil, Argentina, and Mexico, which are in the fourth, sixth, and eighth positions in the collaboration list, respectively, as seen in Figure 2B.
The SJR of the journals in which the papers were published was evaluated as a common indicator of the quality of the journals used.Of all the journals, 85.6% had an SJR and 14.4% did not have an SJR.The median SJR of the papers was 0.789 (rank, 0.000-11.563).The role of foreign international collaboration in publishing in journals with higher SJRs was also evaluated.As can be seen in Figure 3, the SJR was significantly higher (p < 0.0001) in papers that included foreign international collaboration than in papers that did not.The median SJR in the collaborative papers was 1.372 (rank, 0.000-11.563),compared to the SJR of papers without such cooperation (median, 0.486) (rank, 0.000-5.720).The Tukey boxplot shows the median SJR of published papers published with the inclusion of international collaboration.The asterisks indicate statistically significant differences (***p < 0.0001).).Those five categories covered more than 65% of the total scientific production.Table 2 shows the selected journal categories used.
The main virology journal used by Colombian researchers was also investigated.As seen in Table 3, Colombian papers were published in 28 of the world's virology journals.The five journals most frequently used by the Colombian researchers were Virology Journal (10.4%),AIDS Research and Human Retroviruses (9.4%), Archives of Virology (7.3%), Journal of Clinical Virology (7.3%), and Virology (6.3%).However, it is important to highlight that virology journals were only the second category of journals used.
According to the scope of the SJR, virology is one of the widest areas of research.It includes all aspects of viral organisms and host-virus interactions and covers the molecular, biochemical, and cellular studies of bacterial-, archaea-, plant-, animal-, and humanspecific viruses, as well as materials on medical virology and pathogenesis and treatment of viral diseases.In Colombia, various researchers work on viral diseases and viruses.The numbers of papers dealing with each virus or group of viruses according to similar characteristics were analyzed.The major viruses or diseases that appeared in databases were HIV/AIDS, dengue, and papillomavirus (21.4%, 15.8%, and 12.7% of the total productivity, respectively).Plant viruses and some veterinary viruses appeared as the fourth and fifth groups of viruses studied (8.3% and 6.9%, respectively).As can be seen in Table 4, some categories included more than one viral agent.It is important to highlight that research on human viral infections accounts for the vast majority of the papers produced in Colombia (84.5%).
Table 6 shows the top 10 most prolific authors with their affiliations, taking in account the standard competition ranking.These 10 authors accounted for 26.3% of the virology-related published papers in the studied period, and all of them belonged to some of the top 10 institutions (Table 5).A list of the top 10 most-cited Colombian virology papers is presented in Table 7.It is important to highlight that the most-cited papers (8 of 10) focused on HPV, included at least one of the most prolific authors (N.Muñoz), and belonged to one of the most prolific institutions of the country (Instituto Nacional de Cancerología), although none of those top virology papers were published in virology subject journals.

Discussion
This study shows an interesting analysis of Colombian's research publications on viruses and viral diseases between 2000 and 2013.The main results show a continuous increase and important growing trends in the number of publications from Colombia on the subject of virology (Figure 1, Table 1).This is consistent with data from the only previous Colombian bibliometric study [9] that showed an increasing trend in the number of publications on infectious diseases in the main Colombian journals from 2000 to 2009.Using a similar approach, a growing pattern has been shown for Colombia's academic production in the field of biotechnology for the period 2001-2012 [22].However, in contrast to this work, our study analyzed publications from the four main databases and the main international journals to have a better view of the current situation of research on viruses in Colombia.As has been reported previously [23], there are some areas in which a literature search in one or two databases alone leads to incomplete information retrieval about the region, due mainly to the wide scope of the research topic; this issue occurs also for virology.Our data set collected major virology's Colombian papers indexed in the most important regional (SciELO and LILACS) and international (Scopus and PubMed) databases.Falagas et al., in 2005, showed that developing areas of the world (such as Colombia) urgently need more help from developed regions to enhance their research infrastructure in the field of virology [8].This 1235 study found that Colombian researchers are increasingly producing internationally co-authored papers with an increasing number of collaborations (Figure 2A), and the vast majority of these with developed countries such as the United States, Spain, and France (Figure 2B).These first-world countries have had a significant growing share in the virology field in the last decade [4,24], which can be very useful for developing countries such as Colombia in terms in terms of reduction of transactional cost of technology transfer, communication, and quality of research.Important Latin American countries such as Brazil and Mexico are included in the top 10 collaborators, and those countries have been shown to be the most important producers of regional literature in public health registered in Scopus and to act as proxies in the network of intra-regional collaboration in different areas [25,12,11].
One of the most the most important findings of our analysis is that developing countries (in our case, Colombia) could obtain better visibility of their research by developing international projects and internationally co-authored publications (Figure 3).Our results and previous studies [26,27] found that articles with international collaboration appeared in journals with higher SJRs and were likely to be more frequently cited [16].It has been suggested that this may be due to the fact that research involving international cooperation could include topics of broader interest and be suitable for journals with higher SJRs, that only the strongest researchers will have the resources and motivation to overcome the difficulties of collaboration over a distance, and that the biggest-and highest-profile projects are more likely to be international [16].In fact, countries with the lowest scientific productivity (such as Colombia) have a higher tendency to collaborate than do the countries with the highest leadership.Advanced scientific development and leadership may mean greater autonomy and a lesser need to collaborate [28].
Our results and the results of others [29,24,7] have shown that it is important to take into account the fact that many articles concerning viruses and viral diseases are published in journals of other SJR categories and not just in the virology category.This study shows that, for Colombia, the main categories included medicine (miscellaneous), virology, infectious diseases, and public health, environmental and occupational health (Table 2).This could be due to the strategic, tropical, and geographic location of Colombia and the possibility of conducting research on important tropical pathogens such as the dengue virus (Table 3).It should also be stressed that research in virology and/or viral diseases actually interests physicians of several medical specialties and researchers from other biomedical areas such as veterinarians, biologists, agricultural engineers, and others [8,1].
Although Richardson recently (2014) proposed that global bibliometric analysis in virology including only the journals that are categorized in Scopus within the virology subject captures the great majority of relevant research [30], our results shows that Colombian researchers published virology papers most frequently in journals of medicine (miscellaneous) (26%) than in journals of virology (14.1%), showing that it is very important to evaluate different databases [23] and different categories to have a full panorama of what is been published in an area of research.This idea has also been supported by recent studies showing that the global research on yellow fever virus is published mainly in the public, environmental, and occupational health category in the database [31], whereas research on human papillomavirus (HPV) is published in journals in which oncology is the main category [32]; in both cases, virology was the fourth subject in order of importance, confirming our hypothesis that not all virology knowledge is published in journals of the same subject.
Another important finding is the fact that HIV/AIDS, dengue, and papillomavirus research account for almost half of the Colombian articles published.This is very important because those viruses are well known as major pandemic and emerging viruses worldwide [33,34,35], and developing countries need to be ready to fight these and other viral threats.Such results could indicate that Colombian researchers are prepared to face various viral threats as previously encountered in the 2009 influenza A(H1N1) pandemic [36].
One of the most controversial issues in bibliometric evaluation is the use of the SJR to qualify authors' work; one of the limitations of using the SJR is that it is greatly dependent on the category of the subject [37].Our analyses combining SJR allowed us to summarize the SJR of the papers belonging from a same journal categories (Table 2) or a virus studied (table 4) as an arbitrary measure of the visibility of the research in that category and/or virus.In the case of Colombian productivity in research on HPV, although most of the published papers came from HIV studies, the highest median SJR of the publications came from HPV research (Table 4), indicating that the classical view of the number of papers is not the only important 1236 factor in categorizing research.This could be useful in establishing points of reference and facilitating decision making regarding research policies related to virology studies and also animal, human, and plant health, as has been proposed by others [38].
Potential limitations of the present study are related to the databases used to retrieve articles.The databases analyzed do not represent all scientific and biomedical journals published [39].However, this probably does not add any systematic bias in the analysis of the data, and the fact that we used four different international databases diminishes bias.Also, despite the meticulous search methodology, some articles were missed because the full address of the authors was not registered or the address was misspelled.Other limitations include the multiple citations of the subject category in the journals examined that were generally used in SCImago Journal and Country Rank.One journal with its articles might be indexed with several subject categories.However, this fact probably does not add any systematic bias to the analysis of the data because we selected the most accurate subject to the virology knowledge.

Conclusions
This work has revealed a progressive increase in the scientific productivity, visibility, and impact of published research on the subject of virology in Colombia.This is due to the growing presence of publications in foreign journals and increasing international collaboration with developed countries, both of which have allowed better publications in quality journals and better visibility of the virology studies carried out in Colombia.

Figure 2 .
Figure 2. International collaboration in Colombian papers

Figure 3 .
Figure 3. Distribution of the impact factor according to the international collaboration.

Table 1 .
Distribution of the number of authors per publication

Table 2 .
Main journal categories of Colombian papers Standard Competition Ranking according to the number of papers published papers published during the period studied (2000-2013); b The SJR value in the table is the average of the SJR of each one of the journals belonging to each category, according to the SCImago Journal & Country Rank in 2013. a

Table 3 .
Main virology journal used by Colombian researchers a Standard Competition Ranking according to the number of papers published during the period studied (2000-2013).b The SJR was reported according to the SCImago Journal & Country Rank in 2013.

Table 4 .
Principal virus studied in the Colombian papers published a Standard Competition Ranking according to the number of papers published during the period studied (2000-2013); b The SJR value in the table is the average of the SJR of each one of the journals belonging to each category, according to the SCImago Journal & Country Rank in 2013.

Table 6 .
Top 10 most prolific authors with their affiliations