SEARCH

SEARCH BY CITATION

Keywords:

  • citation analysis;
  • bibliometrics;
  • historical article;
  • periodicals;
  • publishing

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methodology
  5. Results
  6. Discussion
  7. References

We examine the bibliometric profile of Tropical Medicine and International Health using the subjects of the articles published and the geographical distribution of the authors. The most common subject areas of papers published during 1996–2003 are highlighted, and the most cited papers indicated.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methodology
  5. Results
  6. Discussion
  7. References

In the final issue of the first volume of Tropical Medicine and International Health (TMIH) , Schoonbaert and Roelants (1996) gave an excellent overview of the use of citation analysis and impact factors, as an assessment tool for the quality of scholarly journals and the research published within them. Bibliometric analysis can indicate trends and patterns within scientific disciplines, national and international strengths and biases in areas of research (May 1997, 1998). There are many ways with which bibliometric and citation tools can be used, and some are more controversial than others. Comparisons between scientists using publication output or impact factors when offering faculty or research positions have often been a point of contention (Calza & Garbisa 1995). Comparing scientists between disciplines is impossible using these tools due to the varying citation rates observed in different fields of research. Bibliometric analysis can be used to take a macro-view of research output, these comparisons and analysis become more meaningful when moving away from individuals towards institutions, nations and geographical regions.

The main tools for bibliometric analysis remain the Institute for Scientific Information (ISI) database, Science Citation Index (SCI Expanded 1945-present), and in tropical medicine, Medline/PubMed, EMBASE, CAB Abstracts and BIOSIS. In the past, Tropical Diseases Bulletin and Current Contents have been used to examine the published literature in tropical medicine. Each of these databases has advantages and drawbacks when conducting bibliometric research. ISI is the most commonly used bibliometric tool because of its indexing of address information for all authors, and citation information for all entries in its database. The main drawback with ISI is the fact that it does not allow for controlled vocabulary search tools such as the Medical Subject Headings (MeSH) used in PubMed/Medline, and the Elsevier thesaurus (EMTREE) used in EMBASE.

The result of this is that free text searching and keywords appearing in the title and abstract fields retrieve articles. The main bias with all these databases is that they are dominated by North American and European publications, and by default, authors. A further bias is their inaccessibility to authors from developing countries. ISI, Medline, EMBASE and BIOSIS are subscription based, and whilst PubMed is freely available to all with an Internet browser, in many countries this is still a formidable barrier to overcome.

The objectives of this study were to look at the bibliometric profile of TMIH and examine the subject content in relation to disease coverage and geographical distribution of the authorship of papers published in the first eight volumes between 1996 and 2003.

Methodology

  1. Top of page
  2. Summary
  3. Introduction
  4. Methodology
  5. Results
  6. Discussion
  7. References

The data for the study was identified by searching in PubMed and ISI Science Citation Index on 31 March 2004. For each disease covered in the study, appropriate keywords were identified using MeSH terms and entry terms identified from PubMed. These terms were then used on the ISI Science Citation Index to find articles published in the journal between 1996 and 2003. Article type was restricted using the ISI limit ‘article’ in publication type. The results were downloaded into a database. Using analytical tools within the database, the data was resolved to geographical representation of authorship using the following regional categories: United States of America (USA), America (excluding USA), United Kingdom (UK), Europe (excluding UK), Africa, Asia and Australasia. Countries were assigned regions on a geographical basis. Middle East countries are counted within Asia, and Egypt is included in Africa. Papers with multiple authorship may be counted more than once when authorship crossed regional boundaries. Citation information was gathered in association with the geographical information. This citation data was used as a bibliometric tool to indicate the strength of the research output published within the journal by regional authorship.

Searching within the data, we also looked at the bibliometric profile of diseases covered by articles within the journal. As part of this analysis, papers covering multiple disease topics may be counted more than once. Discrete counting of papers in the analysis was chosen as a methodology for both the geographical distribution and analysis by disease as proportioning authorship on multiple-authored papers would not accurately reflect the contribution of each author as ascertained by the published paper.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methodology
  5. Results
  6. Discussion
  7. References

Table 1 shows the geographical distribution of authorship for articles published between 1996 and 2003. Europe, with 564 papers, has the largest number of authored papers, which may be expected in a European journal. The other regions display some interesting statistics. Africa has the second highest representation with 517 authored papers. This may reflect the high burden of tropical diseases in the region in addition to the Sub-Saharan epidemic incidence of HIV/AIDS and tuberculosis. The USA, the most prodigious producer of scientific papers, had authorship of only 158 articles. The UK and Asia each had a similar numbers of papers published. When we look at citations to authored papers, African-authored papers have been cited most with 3512 citations.

Table 1.  Geographical distribution of authorship and citations
RegionNo. papers*% TPNo. citations*% TC
  1. * Total papers (TP) is the total number of papers in the study. Total citations (TC) is the total number of citations in the study. Papers with multiple authorship may be counted more than once.

Europe56450.9295648.54
Africa51746.66351257.67
UK26523.92214135.16
Asia25723.19109217.93
USA15814.2681513.38
America11610.475839.57
Australia292.621161.9
Total papers*/ total citations1108 6090 

Table 2 shows the ratio of citations to papers. When working out the relative impact of papers published by region, papers featuring UK authors had a mean citation rate of 8.079. Africa is again the second highest with a citation rate of 6.793. Europe, USA, and America have similar rates ranging between 5.025 and 5.158. The mean for all 1108 papers was 5.496.

Table 2.  Citations per paper (CP) by region
RegionNo. papersNo. citationsCP
UK26521418.079
Africa51735126.793
Europe56429565.241
USA1588155.158
America1165835.025
Asia25710924.249
Australasia291164.000

Table 3 shows the number of papers by disease. Malaria was the most common subject area covered by articles published in the journal, 329 papers were published in the study period 1996–2003. HIV/AIDS and schistosomiasis each had 83 papers. Not all diseases were included and many papers featured more than a single disease; these were counted more than once. Many papers concerned public health, health promotion and policy. These papers are not included in Table 3.

Table 3.  Papers published by disease*
DiseaseNo. papers
  1. * Not all diseases are listed in the table. Papers covering multiple diseases may be counted more than once.

Malaria329
HIV/AIDS83
Schistosomiasis83
Filariasis61
Onchocerciasis48
Tuberculosis47
Leishmaniasis44
Tryanosmiasis38
Diarrhoea22
Dengue21
Typhus21
Meningitis19
Hepatitis19
Measles15
Syphilis11
Tetanus11
Chagas10
E. coli10
Salmonella9
Influenza8
Fascioliasis7
Leprosy7
Diabetes7
Poliomyelitis7
Trachoma7
Typhoid6
Yellow fever5
Ebola4
Rubella4
Cholera3
Gonorrhoea3
Japanese encephalitis3
Plague2
Rabies2

Table 4 shows the five most cited papers in TMIH. The two most cited papers were about insecticide-treated bednets and malaria prevention and control. Both were published in 1996.

Table 4.  List of highest cited articles 1996–2003
RatingArticleNo. citations
1Nevill et al. (1996)150
2Binka et al. (1996)139
3Taylor et al. (1998)62
4Verhoeff et al. (1999)60
5Muller et al. (1996)49

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methodology
  5. Results
  6. Discussion
  7. References

With a journal impact factor in 2003 of 2.156, placing it first within the ISI category Tropical Medicine (ISI 2003), TMIH is a global forum for the dissemination of research and communication in tropical diseases and related issues. The results of the study of geographical distribution of authors demonstrate that the journal reflects fairly the research and views of scientists from all regions of the world. This global distribution would seldom be replicated in the top journals from other ISI categories. These are dominated by journals publishing papers primarily from authors in developed countries. The countries of the world bearing the greatest burden of disease are situated in Africa, South-east Asia, and South America. All these regions were well represented in the papers published during the time scale of the study. Kofi Annan called on the world's scientists to close the research and development gap between developed and developing countries (Annan 2003). TMIH is embracing these challenges by providing a fair and unbiased forum for authors from the developing world to publish their research. More than 500 papers were authored or co-authored by African researchers.

Malaria was the subject area with the largest number of publications, over four times as many papers as were published for HIV/AIDS or schistosomiasis. Tuberculosis saw 47 papers published. One may surmise that with the increase in incidence of tuberculosis caused by co-infection associated with HIV/AIDS, we will see a rise in the number of papers published on both these subjects over the next 8 years, as well as on the chemotherapeutic interventions that need to be developed to combat the epidemics. During the period of study there were seven papers published in the Journal about poliomyelitis. If the World Health Organization eradication programme finally delivers its objectives, then maybe we will see the last of these papers in the coming years.

Citation rates were strong for papers with UK authors. Citation analysis has many biases. This study used the level playing field of using papers from within a single journal. The UK reservoir of papers was smaller than both Africa and Europe, but the average citation per paper was higher. Finally, two papers have been cited more than 100 times. Other studies have called papers with 100 plus citations as ‘citation classics’ (Baltussen & Kindler 2004), although different figures have been used to indicate a citation classic. These two papers, plus the fifth most cited paper were published in 1996. The cited half-life for TMIH was 3.8 years (ISI, Journal Citation Reports 2002), this indicates the average time an article published in TMIH takes to amass half of its eventual citations. Given the cited half-life it could be expected to see older papers appearing towards the top of the list. The paper by Verhoeff et al. (1999) could be expected, if applying the cited half-life principle, to gather close to 91 citations by 2006.

Bibliometric analysis using citation data is not an exact science. All citations are counted by the ISI, even if they are self-citations or negative citations. Self-citations may well be justified if authors are referring to prior work of importance to the citing article. Negative citations cannot be distinguished when using the Science Citation Index and any bibliometric study would have to assume that self-citations and negative would even-out over time.

The mean citation rate of the 1108 papers was 5.496. Four papers were cited over 50 times, 362 papers had received either 1 or no citations. This is consistent with Seglen's findings that a small number of highly cited papers have a disproportionate effect on a journal's impact factor (Seglen 1992, 1994). The two papers with over 100 citations both concerned the use of insecticide-treated bednets in reducing the incidence and mortality of malaria in child populations.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methodology
  5. Results
  6. Discussion
  7. References