16 A Multi-Level Analysis of World Scientific Output in Pharmacology

Over the last few decades and particularly in the present economic context, the distribution of economic resources has been a concern addressed by governmental and corporate scientific policy, which has either benefitted only part of the scientific and technological community or furthered certain lines of research. The pharmaceutical industry in particular has had to confront not only this situation, but also ongoing internationalisation, supported by the relentless advances in communication technologies.


Introduction
Over the last few decades and particularly in the present economic context, the distribution of economic resources has been a concern addressed by governmental and corporate scientific policy, which has either benefitted only part of the scientific and technological community or furthered certain lines of research.The pharmaceutical industry in particular has had to confront not only this situation, but also ongoing internationalisation, supported by the relentless advances in communication technologies.
Until the nineteen eighties, industry internationalisation, in terms of R&D, was a marginal matter, not only for economics theory and business in general, but also for governments and the other organisations involved.Globalisation began to acquire importance after the mid nineteen nineties, although not all manufacturing industries have experienced the same degree of R&D internationalisation.The pharmaceutical industry, for one, pioneered this more universal approach to research and development (Noisi, 1999).
Contrary to the widely held opinion according to which R&D internatianlisation is the fruit of domestic innovation in many industries, pharmaceutical constitutes an exception.Indeed, international innovation intensifies the industry's R&D (Patel and Pavitt, 2000), whereas in other lines of business domestic innovation is the driver.In addition to internationalising its R&D, the pharmaceuticals industry has increased its research spending exponentially in recent years (Congressional Budget Office, 2006).
A number of earlier papers studied the bibilometric characteristics of the pharmacological publications generated as a result of the R&D effort in places such as the United States (Narin and Rozek, 1988), India (Kaur and Gupta, 2009;Gupta and Kaur, 2009) or the Middle East (Biglu and Omidi, 2010).Others stressed the contribution of pharmaceutical firms to scientific knowledge (Koening, 1983;McMillan and Hamilton, 2000;Rafols, et al. 2010;Perianes-Rodríguez, et al. 2011).The assessment of the international impact of scientific papers is a present, but not a new concern: it has been a frequent object of study since the nineteen eighties.The use of scientific indicators for several decades to characterise research by subject area, country or institution has confirmed that, although they have their limitations, they are the only suitable tool for scientific assessment (Braun T et al., 1985).
A Multi-Level Analysis of World Scientific Output in Pharmacology 341 When papers were co-authored by researchers from institutions in different countries, a complete computational approach was adopted.The growth rate, when provided, indicates the rise or decline in world-wide output in 2009 with respect to the baseline year, 1996.
A number of indicators were used to obtain an approximate view of the quality of world scientific output in the field of pharmacology.The number of citations received refers to the total number of times papers published by the unit analysed were cited during the period studied.This indicator provides an overview of the scientific impact of the articles published by the unit in question.The number of citations per paper was calculated as the mean number of citations received by all the papers published by the unit analysed in the period studied.
The domestic citations were separated from the total to determine the proportion of the output that was used as a reference in the same geographic area (region or country) and consequently, by simple subtraction, the proportion involving knowledge transfer to other areas.The results are shown as the percentage of the citations used for research conducted in the same geographic area.The normalised citation indicator is the relative number of times papers produced by a specific unit were cited, compared to the world-wide mean for papers of the same type, age and subject area.
While citations denote the subsequent use of papers once published, the references list the literature cited in papers published by a journal at any given time.The number of references per paper was found by dividing the total number of references by the number of papers published by the unit.A country's H-index, in turn, specifies the number of papers (h) produced in that country and receiving at least h citation.It relates a country's scientific productivity (output) to its scientific impact (citations).The international collaboration indicator is the percentage of papers with author affiliations in more than one country.This indicator measures institutions' international networking capacity.In this chapter a journal's % output in Q1 is the percentage of scientific papers published by an institution in what are classified as the most influential journals in the respective category, i.e., the periodicals in the first quartile or Q1, the upper 25 %, based on their SJR value.
Another qualitative indicator used, homonymous with the aforementioned scientific information system (SCImago Journal and Country Rank), was the Scimago Journal Rank (SJR), used as an alternative to the traditional impact factor (I.F.).This indicator, which measures the visibility of the journals in the Scopus® database, is established by the SCImago2 research team on the grounds of the well-known Google PageRank TM algorithm.It differs from the I.F. in two ways: citations are computed over 3 rather than 2 years; and article citations are weighted, with citations in more visible or prominent journals carrying greater weight than citations in lower-ranking journals (González-Pereira et al., 2009).

World-wide science and pharmacology
World-wide scientific output, as listed in the Scopus database for the period running from 1996 to 2009, came to 21 100 138 papers.The total citations received by those papers during the same period amounted to 217 388 448, for a mean of 10.03 citations per paper.The absolute numbers for pharmacology, as one of the 27 subject areas established by Scopus, were logically much smaller.The totals were 564 914 papers and 6 266 408 citations.The mean number of citations in pharmacology was therefore higher than the world average, at 11.09.The growth rate for this subject area was 4.76 %, reflecting the growth in its scientific output.
Figure 1 shows the percentage contribution of the Scopus subject areas to world-wide scientific output during the period studied.Medicine played a predominant role in the international scientific scenario, with a mean yearly contribution of over 20 %.Decision science and dentistry stood at the other extreme, with a mean yearly output of 0.35 %, shown on the figure as very thin lines.The mean yearly contribution of pharmacology to international scientific output in the period was 2.7 %, shown in red on the right half of the graph.When pooled, all the subject areas with relative outputs of under 4 %, which include pharmacology, earth and planet sciences, immunology and microbiology, accounted for 34.83 % of the scientific papers published world-wide.Fig. 1.World output by subject areas (%) (Scopus, 1996(Scopus, -2009) )

Pharmacology by region
W h i l e s c i e n t i f i c o u t p u t b y r e g i o n i s a n important indicator to determine regional contributions to pharmacology, quantitative information alone is incomplete and must be supplemented with data on the impact of these papers on the scientific community.Table 1 gives the vales of some of the indicators described earlier for a number of regions, along with colour bar graphs for readier interpretation.(Scopus, 1996(Scopus, -2009) ) The behaviour of the domestic citations indicator merits comment.In North America, these citations accounted for over 80 % percent of the total.The number of domestic citations was likewise very high in Western Europe; in both regions most of the citations were found in articles published in the same country as the paper cited.Consequently, in these two regions, the large number of domestic citations led to an inordinately large number of total citations.
The regions with smaller numbers of citations also had a smaller proportion of domestic citations.In other words, their output was acknowledged primarily by other regions, while domestic citations were less frequent.The region that best illustrates this observation is Northern Africa, where only 23.92 % of the citations received were domestic.
The number of citations per paper was also highest in North America and Western Europe, with the Pacific Region ranking a close third.Central Africa's low scientific output in pharmacology was only scantly acknowledged, with only 5.46 citations per paper on average.Asia, Eastern Europe, Latin America and Northern Africa had similar citations per paper values, which ranged from 8 to 9.
The pharmacological output by regions over the period 1996 to 2009 is shown in Figure 2. The three most productive regions in that period were Western Europe (red), North America (blue) and Asia (green).Asia had a higher growth rate in the latter years of the period and was the most productive region in 2009.This rise may have been the result of greater participation in the pharmacology, particularly in countries that in those years began to adopt a very active role in the field.

Countries and pharmacology
The basic unit for the regions listed above was defined as the individual country.A total of 194 countries published pharmacological research in the period studied.The analysis conducted of their output provided greater insight into the values found for the regional indicators.
Fig. 2. Pharmacological scientific output by region (Scopus, 1996(Scopus, -2009) ) The ten most productive countries accounted for around 71 % of world-wide pharmacological output in the period studied.These ten countries are listed in Table 2, which shows their total output in the period, the number of total and domestic citations received, the citations per paper and the H index.The list is headed by the United States, which had the largest output and number of citations, although the number per paper should be interpreted bearing in mind the impact of the large number of domestic citations identified.At 293, its H index was likewise high, indicating that 293 papers were cited in 293 other articles.
Table 3 ranks the countries whose overall data for the entire period are given in Table 2, year by year across the period.Grey shading indicates that the country changed its position from the preceding year and maroon shading that the country joined the top ten in the year in question.
The regional study showed the enormous progress in Asia in the latter years of the period.That growth was the result of greater participation in the subject area by Asian countries.
Although until 2005 Japan was the second largest producer in pharmacology, from 2006 onward it was overtaken by an emerging neighbour: China.In the three earliest years China ranked tenth; in the intermediate years it gradually climbed to higher positions and finally reached second place in 2006.While still among the most productive countries, Japan's position slid, denoting its tendency to contribute less and less to pharmacological output.In the last year of the series, 2009, four of the ten most productive countries were Asian (China, India, Japan and South Korea).
The United States maintained its lead throughout the period.That leadership and Canada's contribution, from lower but still productive positions, made North America the sole region with an output comparable to Asia's in the latter years.All the other most productive countries in pharmacology were from Western Europe: United Kingdom, Germany, Italy and France, and the Netherlands and Spain in some years.Only one Latin American country was among the most productive during the period: Brazil, in 2007.(Scopus, 1996(Scopus, -2009) ) Figures 3 and 4 show the relationship between international collaboration and citations per paper in countries publishing at least 1 000 papers.The position occupied by the countries in each region is shown in both figures, but only Western European and North American countries are depicted in Figure 3.All the Asian, Eastern European and Latin American countries are shown in Figure 4, although only the BRIC countries (Brazil, Russia, India, China) are labelled.The country in Figure 3 with the smallest number of citations per paper and least intense international collaboration was Turkey.With 6.18 citations per paper and an international co-authorship percentage of 16.12, it stood at the low end of its region, Western Europe, and had lower citation values than Latin America or Eastern Europe.In Western Europe, Sweden and Belgium were the two countries both in that region and the world with the highest international collaboration indices and a mean of 12 citations per paper.Both, as well as other countries, also had higher values than the USA (in terms of international collaboration) and Canada.
Measuring their scientific status in terms of citations per paper and international collaboration values, the BRIC countries still have room for improvement.Three of those four countries were positioned very close to the origin on the graph.Of the four, only Brazil showed values close to the results recorded for Turkey.

Pharmacology in journals
The analysis of the journals that published pharmacological papers included the data for the periodicals that published at least one such paper in 2009.Under that criterion, a total of 482 journals were identified, 61 of which had been recently added to the database and consequently lacked the data needed to calculate their SJR.
The remaining journals were published in a total of 33 countries, each with less than six journals.The large and unwieldy original table was abbreviated to build Table 4, which gives the values for only the journals with the 10 highest SJR and the 10 scientific journals that published the largest number of pharmacological articles in the last year of the series.Note that none of these journals appears on both lists.
Of the scientific journals with the highest SJR, two were published in the US, Annual Review of Pharmacology and Toxicology and Pharmacological Review, and one in the United Kingdom, Nature Reviews and Drug Discovery.These three journals had SJR scores of 3.56, 3.3 and 2.68, respectively.That means that they received large numbers of citations, but also that since they are weighted by journal prestige to calculate the indicator, those citations appeared in other high quality journals.Neither of the US journals was very productive, with only 19 papers each in 2009, compared to a much larger output by the English periodical, which published a total of 202 articles.
The scientific journals with the highest output in pharmacology were The Netherlands' Bioorganic and Medicinal Chemistry Letters, with 1 546 papers, followed by the UK's Pharmaceutical Journal, with 1 058 and Germany's Deutsche Apotheker Zeitung, with 967.Their SJR indices were lower than for the journals mentioned in the preceding paragraph, however, with scores of 0.21, 0.03 and 0.02, respectively.In other words, in the period calculated for the SJR index (three years), either the absolute number of citations received by this group of more productive journals was very low or the citations were published in lower quality journals.
Each country's contribution to pharmacological scientific output can be analysed from two perspectives: as specified earlier, by the contribution made by its scientists through their published papers, or by the journals edited in the country.These two factors are compared in Figure 5.Each country's scientific output is shown in red and its publishing activity in blue.Many countries, such as the United States, show similar percentages for both types of contribution, while in others the values vary widely.A case in point is The Netherlands, whose scientific output was a mere 2 % while its journals published over 20 % of the pharmacological articles.www.intechopen.com

Scientometric indicators for pharmaceutical companies
The pharmaceutical industry, in addition to being one of the most profitable, is also one of the most globalised and fastest growing lines of business.Moreover, its large investment in research makes it an innovation-intensive activity.This innovation is the result of the direct or indirect interaction of a large number of actors: different types of companies, research institutes, financial institutions, public bodies and authorities, public and private universities, research centres, regulating bodies, governments, health systems, consumers and physicians, to name a few.
The industry comprises three categories of companies.The first covers (primarily North American and European) multinational companies that operate globally and invest huge sums in R&D, which is centralised in some cases and decentralised with laboratories in many countries and on many continents in others.The second category consists of small companies that supply their domestic markets with drugs that require no substantial R&D investment.The third includes firms that specialise in biotechnology and invest considerable sums in research despite their small size.
In 2010, biopharmaceutical companies invested an estimated 67.4 billion dollars in pursuit of new drugs (Figure 6).The total R&D spending by Pharmaceutical Research and Manufacturers of America (PhRMA) members, including industry majors such as AstraZeneca, Bayer, Boehringer, Ingelheim, Bristol-Myers, Squibb, Eli Lilly, Genzyme, GlaxoSmithKline, Hoffmann-La Roche, Merck, Novartis, Pfizer, and Sanofi-Aventis, as well as non-members, are shown in the figure.1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 1975-2010(PhRMA, 2011) ) Bibliometric indicators can be constructed for the pharmaceutical industry on the grounds of the research results made public by the authors.As noted earlier, the industry has been gradually internationalising its high research and innovation potential since the mid nineteen seventies (McMillan and Hamilton, 2000).
The values of the bibliometric parameters for the pharmaceutical majors are given in Table 5.The data, which cover a seven-year period and are based on these companies' research publications, reveal a number of interesting differential characteristics.The ranking criterion followed was scientific output defined as the number of papers published in 2003-2009, initially disaggregated, although some of the companies listed had parent-subsidiary relationships.
The first significant result was the volume of scientific papers published by these companies.These elite, all of whose members published at least 125 papers in the period considered, was headed by the Pfizer headquarters site, which averaged 353 papers yearly throughout the period, followed by Merck with a yearly mean of 251., 2003-2009 (www.scimagoir.com)Pharmaceuticals is generally agreed to be one of the industries whose research is most intensely internationalised, defining that to mean the proportion of the research conducted outside the headquarters country.The industry's business has become more international since the nineteen nineties as a result of the convergence of a number of processes.New industrial activities have cropped up around biotechnological research, primarily in the US;

Table 1 .
Pharmacological scientific output, citations and domestic citations by region

Table 3 .
Country position by output

2009 2010 Spending (Billions of Dolars $)
Figure 7 shows the R&D spending by PhRMA members in and outside the United States.The total R&D investment by pharmaceutical companies has continued to rise.In 2010 PhRMA members invested 49.4 billion dollars, up 6 % from 2009 and 90 % since 2009.

Table 5 .
The second statistic of interest was the citations per paper, which ranged from fairly low (7.86 for Dow Chemical Co., 8.47 for the Indian firm Dr Reddy´s and 9.26 for Sanofi-Aventis GmbH in Germany) to very high values (18.47 for Astra Zeneca in the United Kingdom and La Roche in Switzerland).These findings suggest substantial differences in the visibility or quality of firms' scientific knowledge.Bibliometric performance indicators for pharmaceutical firms