Article Text

Download PDFPDF

Bibliometric analysis of military trauma publications: 2000–2016
  1. Mark L Vickers1,
  2. C P Coorey1,
  3. G J Milinovich1,
  4. L Eriksson2,
  5. M Assoum3 and
  6. M C Reade1,4
  1. 1 Faculty of Medicine, The University of Queensland, St Lucia, Queensland, Australia
  2. 2 Herston Health Sciences Library, The University of Queensland, St Lucia, Queensland, Australia
  3. 3 Centre for Child Health Research, The University of Queensland, South Brisbane, Queensland, Australia
  4. 4 Joint Health Command, Australian Defence Force, Canberra, Australia
  1. Correspondence to Mark L Vickers, Faculty of Medicine, The University of Queensland, St Lucia, QLD 4072, Australia; mark.vickers{at}uq.net.au

Abstract

Introduction Bibliometric tools can be used to identify the authors, topics and research institutions that have made the greatest impact in a field of medicine. The aim of this research was to analyse military trauma publications over the last 16 years of armed conflict in order to highlight the most important lessons that have translated into civilian practice and military doctrine as well as identify emerging areas of importance.

Methods A systematic search of research published between January 2000 and December 2016 was conducted using the Thompson Reuters Web of Science database. Both primary evidence and review publications were included. Results were categorised according to relevance and topic and the 30 most cited publications were reviewed in full. The h-index, impact factors, citation counts and citation analysis were used to evaluate results.

Results A plateau in the number of annual publications on military trauma was found, as was a shift away from publications on wound and mortality epidemiology to publications on traumatic brain injury (TBI), neurosurgery or blast injury to the head. Extensive collaboration networks exist between highly contributing authors and institutions, but less collaboration between authors from different countries. The USA produced the majority of recent publications, followed by the UK, Germany and Israel.

Conclusions In recent years, the number of publications on TBI, neurosurgery or blast injury to the head has increased. It is likely that the lessons of recent conflicts will continue to influence civilian medical practice, particularly regarding the long-term effects of blast-related TBI.

  • statistics & research methods
  • trauma management
  • neurological injury

Statistics from Altmetric.com

Key messages

  • Bibliometric analysis of military trauma publications from recent conflicts can identify trends, important findings and key players in this field.

  • This analysis identified a recent plateau in publication output as well as an increase in focus on the topic of traumatic brain injury, neurosurgery or blast injury to the head.

  • High levels of author and institutional collaboration were identified within countries but relatively little collaboration between countries.

Introduction

The terrorist attacks against the USA in 2001 marked the beginning of the US-led International Coalition Against Terrorism and a new era of combat operations in the Middle East and Central Asia. The ensuing wars in Afghanistan and Iraq led to major paradigm shifts in military trauma care, many of which altered medical practice outside of the war zone.1–3 In particular, experience in these conflicts changed the approach to damage control resuscitation, blood component use, techniques in temporary vascular shunting, preperitoneal pelvic packing and serial evaluation and debridement of traumatic amputations in the patient with polytrauma.3 4 Those involved in prehospital trauma care also learnt from recent military experience with arterial windlass tourniquets, intraosseous catheter access and haemostatic agents.5 6 Military utilisation of trauma databases demonstrated these to be powerful resources for both research and quality assurance.7 8 As the armed forces of much of the developed world return to contingency operations, it might be useful to identify those aspects of military trauma care that have had the greatest impact, in order to facilitate translation into civilian medicine and enduring military doctrine.

Bibliometrics uses quantitative measures to evaluate academic productivity and remains the best approach to analysing the publications of an academic discipline.9–11 In recent years, trauma medicine has employed bibliometric methods to identify articles and authors of influence.12 13 Two of the most widely adopted measurements of author and journal research impact are the h-index and journal impact factor. H-index is defined as the number of papers (h) that have a minimum h number of citations, while the journal impact factor is typically formulated based on the citations recorded over 1 year for papers published in the journal during the preceding 2 years.9 11

There has been no recent comprehensive bibliometric analysis of publications relevant to military trauma. Therefore, the aim of this research was to use bibliometric indices to identify the military trauma publications from the last 16 years that have had the greatest clinical and scientific impact. In addition, the research sought to identify the publication trends as well as authors, institutions and journals who have been the greatest contributors in this field.

Methods

Publications were identified through a literature search of the Thompson Reuters Web of Science database. A single author manually reviewed and categorised 5496 publications by title and abstract to include only publications that dealt with civilian and/or military casualties of war who suffered severe physical trauma arising during any military conflict. Categories were devised through consultation with senior authors in the field. Publications were included if they related to any aspect of casualty care including but not limited to resuscitation, surgery, intensive care, infection control, evacuation, rehabilitation and long-term sequelae. Search strategy and topic categories are provided in Figure 1. No language restrictions were applied. The literature search was conducted on 18 December 2016.

Figure 1

Consort diagram of search strategy and publication categorisation.

Web of Science tools was used to identify author h-index, mean citation counts, language/country of origin and the 30 most cited publications. Features of the 30 most cited publications reported on included topic area, publication type, study type, publishing journal and highly contributing authors. The 30 most cited publications were reviewed in full. We described the most cited overall, the most cited prospective study and the most cited study conducted in a war zone. To describe the total dataset, publication details were exported to Endnote (Clarivate Analytics, USA), R (R Core Team, Austria) and R packages ggplot214 and Circlize.15 The following features were analysed: document types, author collaborations, topic areas, publishing journals and collaboration between institutions and countries. To distinguish between authors, the all initials method of disambiguation was used, where authors with the same first initial but different middle initials were considered unique. For analysis, the following organisations were combined due to prior amalgamations, staff overlap and/or geographical proximity: (1) Washington DC military institutions (National Navy Medical Centre Bethesda, Uniformed Services University of Health Sciences, Walter Reed Army Institute of Research and Walter Reed National Military Medical Centre); (2) San Antonio military institutions (San Antonio Military Medical Centre, US Army Institute of Surgical Research and Brooke Army Medical Centre); and (3) UK institutions (Royal Centre for Defence Medicine including the Queen Elizabeth Hospital Birmingham). International collaboration was determined to have occurred where a paper listed authors from two or more different countries based on author address and/or affiliated institution as extracted from Web of Science.

Results

Top 30 most cited

The most common topic of the 30 most cited publications was wound and mortality epidemiology with nine documents; the second most common topic was traumatic brain injury (TBI), neurosurgery or blast injury to the head with eight publications followed by four publications each focusing on transfusion, shock, emergency medicine or critical care and haemostatic agents or tourniquets. Eighteen publications (60%) were retrospective chart/registry reviews while only four (13%) were prospective studies. Of the remaining, four were topic/situation overviews, three were editorials and one document was a proceedings paper. The leading journal was the Journal of Trauma and Acute Care Surgery (JTACS) which published 11 (37%) of the 30 most cited papers, followed by the New England Journal of Medicine with five (17%) and the Annals of Surgery and the Journal of Head Trauma Rehabilitation, each with two publications (7%). Four authors from our top 15 contributing authors list were also the primary authors of documents in the 30 most cited list. Eastridge BJ was the primary author of three documents, and Owens BD and Holcomb JB were primary authors of two publications each. Holcomb JB was the last (presumably senior) author of 10 (33%) of the 30 most cited publications. The only other authors to be listed as senior author of more than one paper were Warden D and Blackbourne LH who were the senior authors of two papers each. The 30 most cited articles are listed in Table 1.

Table 1

Top 30 most cited publications in final dataset (n=2132) and number of citations

The most frequently cited article overall was by Borgman et al. ‘The ratio of blood products transfused affects mortality in patients receiving massive transfusions at a combat support hospital’. This publication described a retrospective chart review of 246 massively transfused patients at a North Atlantic Treaty Organisation Role III hospital in Iraq between 2003 and 2005.16 The publication compared mortality in patients receiving plasma and red blood cell transfusions, finding that the lowest mortality was associated with ratios of close to 1:1. This paper was cited 671 times which was 269 more citations than the second most cited publication. The most cited prospective study was by MacDonald et al 17 with 232 citations, ‘Detection of blast-related traumatic brain injury in US military personnel.’ In this publication, the authors described the use of diffusion tensor imaging to detect traumatic axonal injury in soldiers with primary blast exposure evacuated from Iraq or Afghanistan to the Landstuhl Regional Medical Centre between 2008 and 2009.17 Among 63 cases, the authors detected brain abnormalities not typically known to be injured in non-blast TBI. Of the four prospective studies, only one was conducted in a war zone. With 165 citations, this work by Kragh et al 18 surveyed 232 casualties with 428 tourniquets at a combat support hospital in Baghdad over 7 months in 2006. In their patient population, the authors demonstrated that the duration of tourniquet application was not associated with increased complications such as thrombosis, myonecrosis, pain, nerve damage, renal failure, amputation or fasciotomy.18

Total dataset analysis

Our Web of Science search string identified 5496 potentially relevant publications. Manual evaluation and categorisation by title and abstract left 2132 publications in the final dataset. The majority of publications were written in English (n=2072), followed by German (25) and French (19). The USA was the most common country of origin (n=1400), followed by the UK (296), Germany (126) and Israel (82). The final dataset comprised 2075 journal articles, 44 books and 13 book sections.

In total, 5699 unique authors contributed to at least one publication in the dataset. Among the top 15 contributing authors, the two who collaborated most frequently were Holcomb JB and Wade CE with 35 mutual publications. This was followed by Blackbourne LH and Holcomb JB (25). Collaborations between the top 15 contributing authors are shown in Figure 2, along with author dataset h-indices, mean citations per publication and total number of contributing publications. The median number of collaborations for any possible pair of the top 15 contributing authors was 3. The highest contributing author overall was Holcomb JB (100 publications), who was also the author with the highest dataset h-index: 41.

Figure 2

Network of collaborations between the top 15 contributing authors by publication volume to the final dataset (n=2132). The outer rim represents the number of collaborations an author has made with other authors in the top 15 list, with spokes representing collaborations between authors. Only those publications which have been authored by at least two of the top 15 authors are included. 

The most common topic area over the entire study period was transfusion, shock, emergency medicine or critical care (397 documents, 18.6%). This was followed by orthopaedics, amputation and surgery to the limbs or spine (356, 16.7%) and TBI, neurosurgery or blast injury to the head (334, 15.7%). As a percentage of output annually, the topic area that demonstrated the greatest absolute increase from 2000 to 2016 was TBI, neurosurgery or blast injury to the head, which comprised 14% of articles in 2000 and 29% of articles in 2016. The next greatest increase was in transfusion, shock, emergency medicine or critical care which was the topic area of 9% of articles produced in 2000 and 17% of articles produced in 2016. Although the number of publications per year concerned with wound and mortality epidemiology remained stable across the study period, this topic demonstrated the greatest decrease as a percentage of annual volume, comprising 26% of all articles in 2000 and just 5% of publications in 2016.

A total of 377 journals were represented in the final dataset. The journal with the greatest overall contribution was JTACS (previously known as the Journal of Trauma, Injury, Infection and Critical Care) with 345 documents (impact factor 2.80). This was followed by Military Medicine (319, IF 0.97), Journal of the Royal Army Medical Corps (91, IF 0.77), Injury: International Journal of the Care of the Injured (66, IF 1.91) and the Journal of Orthopaedic Trauma (31, IF 1.84). The year with the highest number of publications was 2011 (n=215). The number of publications per year increased steadily from 2000 (n=66) until 2008 (173) and remained relatively stable from 2008 until 2016 with a mean of 181 publications per year during this period. Publication volume by year is shown in Figure 3.

Figure 3

Annual publication volume across the study period.

Of the 2132 publications in our dataset, author organisation affiliations could only be extracted for 2083 publications. The top contributing organisations based on these data were the Washington DC military institutions cluster with 604 documents. This was followed by San Antonio military institutions with 318 publications; UK institutions (152); Madigan Army Medical Centre (88) and the Landstuhl Regional Medical Centre (65). Collaborations between the top 15 contributing organisations and their number of publications are shown in Figure 4. The median number of collaborations for any possible pair of the top 15 contributing organisations was 2. Washington DC military institutions and San Antonio military institutions had the greatest number of collaborations with 94 documents that contained at least one author from each institution. This was followed by authors from San Antonio military institutions and the Madigan Army Medical Centre (30). International collaboration was calculated based on the same data obtained for author organisation. Overall, the USA had the most individual collaboration instances (236), followed by the UK (89), Germany (80) and Netherlands (27). The USA and Germany collaborated the most with 68 publications that contained at least one author from each country. This was followed by the USA with the UK (50), Iraq (16), Canada (11) and Norway (10). International collaboration is depicted in Figure 5.

Figure 4

Network of collaborations between the top 15 contributing organisations by volume of publications in the final dataset (n=2083). The outer rim represents the number of collaborations an institution has made with other institutions in the top 15 list, with spokes representing collaborations between institutions. Only those publications which include authors from at least two of the top 15 organisations are included.

Figure 5

Choropleth depicting publications over total study period (2000–2016) by country of origin and undirected graph of collaborations between countries (n=2083). Countries with three or fewer collaborations are not shown. Lines with number and country name show the total collaborations of that country with the USA. Lines with number only show the total collaborations of the UK with the adjoining country. AFG, Afghanistan; SWE, Sweden; NOR, Norway; DEU, Germany; GBR, United Kingdom; NLD, Netherlands; FRA, France; ITA, Italy; ISR, Israel; IRQ, Iraq; CAN, Canada; AUS, Australia.

Discussion

The volume and focus of military trauma publications have changed in recent years, with a general plateau in the overall number of publications and a notable increase in focus on the topic area of TBI, neurosurgery or blast injury to the head. There were extensive collaboration networks among highly contributing authors and institutions within countries during the study period but low collaboration occurring between authors from different countries.

Changes to the annual volume of publications on military trauma may in part be attributable to changes in coalition military medical activity. Towards the end of the study period, there was both a reduction in numbers of deployed coalition troops and a total reduction in coalition fatality rates.19 It is likely that drop-offs in both fighting intensity and soldier mortality contributed to the general plateau in publications on military trauma. Also, it might be that funding for military trauma research has declined in concert with reducing military commitments, although there is little published commentary to support this contention. Publications on TBI, neurosurgery or blast injury to the head were prominent in our highly cited list and increased as a percentage of annual output across the study period. This finding is in contrast to the results of Orman et al 1 whose 2012 citation analysis reported that damage control resuscitation was the central topic at the time and TBI was an under-represented issue. We suspect that a shift in injury patterns among troops has contributed to this change in publication focus. As International Coalition Against Terrorism partners now experience a peak in the number of returned service personnel, it is likely that the long-term effects of combat exposure, including the sequelae of blast-related TBI, are becoming more relevant. This is not surprising considering that in the USA alone more than 300 000 returned service men and women are currently suspected of suffering blast-related TBI.17 20 21 Other factors contributing to this apparent increase in interest may be the development of new brain diagnostic imaging modalities or a coinciding growing public awareness of sports-related brain injury.22 23 Although beyond the scope of this study, several recent publications are exploring the link between TBI in veterans and psychiatric conditions like posttraumatic-stress disorder (PTSD).24 25

To our knowledge no prior studies have evaluated the extent of author, institution or international collaboration in publications on military trauma. Extensive collaborations between highly cited authors and between institutions within countries were identified but relatively low levels of collaboration between authors from different countries. Strong collaboration networks were likely to be important to maintaining publication productivity across the study period. In a field where the majority of authors are both clinicians and military personnel, it is possible that a number of factors contributed to limited international collaboration including heightened security concerns, language barriers and the transient nature of relationships between integrated clinical service providers in the deployed environment. Although our results showed the highest level of collaboration between the USA and Germany, we expect that a number of authors who were affiliated with German institutions were US personnel and as such, this finding may be biased by an historically strong US military presence at the Landstuhl Regional Medical Centre. We believe that the majority of true international collaboration occurred between the USA and the UK. Because the ability to analyse author networks and their effects on research quality and citation dynamics has only emerged recently with the advent of large digitised publication databases, there is little to compare with in this field.10 Based on our observations, however, we would suggest that there is scope to increase the level of international collaboration when publishing on military trauma.

This study has several limitations. First, documents from non-English speaking countries, in particular China and Russia, appear to be under-represented. Second, while research with animal models was omitted we believe that this may have resulted in the exclusion of relevant primary data, particularly related to TBI. This may have led to the under-representation of several non-US organisations, such as the Defence Science and Technology Laboratory in the UK. Finally, it should also be pointed out that bibliometric tools exhibit a degree of inherent bias. The h-index, for instance, is reliably insensitive to potentially important authors with small sets of highly cited articles and similarly, citation counts may under-represent highly significant but more recent publications due to citation count lag time.9–11

Conclusion

Our analysis revealed that there has been a plateau in the total number of publications on military trauma and a change to the focus of research output. The greatest increase in the percentage of annual publication has occurred in studies focused on TBI, neurosurgery or blast injury to the head and it is likely that the translation of military experience in this field to the civilian medical sphere will be felt further in the coming years. There is also greater collaboration between highly contributing authors and institutions within countries than between authors of different countries.

References

Footnotes

  • Contributors All authors were necessary to completion of this paper. MLV conducted the initial systematic review, co-ordinated all other authors and performed the majority of drafting throughout. CPC conducted all statistical analysis on ’R' and generated the circular graphs and line graph. GJM guided discussion and contributed on multiple drafts of the paper. LE generated the initial search strategy using Web of Science and collated all results with Endnote. AM generated the global map using ARCGIS and bibliometric data. MCR was responsible for devising the initial concept, preliminary analysis and guiding the overall direction of the research paper. MCR also contributed to multiple drafts.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Data kept on the personal computers of the primary researchers.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.