International comparison of regional trauma registries

Injury, Int. J. Care Injured 43 (2012) 1924–1930

Contents lists available at SciVerse ScienceDirect

Injury journal homepage: www.elsevier.com/locate/injury

International comparison of regional trauma registries Hideo Tohira a,b,*, Ian Jacobs a, David Mountain a, Nick Gibson a, Allen Yeo c a

School of Primary, Aboriginal and Rural Health Care, The University of Western Australia, Australia Trauma Registry Committee of the Japanese Association for the Surgery of Trauma, Japan c Department of Surgery, Sir Charles Gairdner Hospital, Australia b

A R T I C L E I N F O

A B S T R A C T

Article history: Accepted 3 August 2012

Background: A regional trauma registry (RTR) collects injury data from multiple hospitals in a given region; however, the differences among RTRs have not yet been thoroughly investigated. Aim: The objective of this study was to identify RTRs worldwide and describe the structural differences, inclusion criteria and demographics among RTRs, as well as to investigate the effect of the inclusion criteria on patient demographics. Materials and methods: We included state, national and multinational trauma registries in this study. We searched for RTRs using the MEDLINE database and a general Internet search engine. We abstracted the funding sources, AIS versions, data submission methods, inclusion criteria and patient demographics of each RTR. We selected the following three outcome measures for comparison: the number of case registrations per year per hospital, proportion of cases with an Injury Severity Score (ISS) > 15 and crude mortality rate. We compared the outcome measures for RTRs that included ‘an ISS > 15’, ‘an admission to the Intensive Care Unit (ICU)’ or ‘a transferred patient for higher care’ with those of RTRs that did not. Results: We identified 17 RTRs (11 national, 4 state and 2 multinational). Government funding was the most common funding source. The RTRs most frequently used the AIS 98 or AIS 2008. Web-based data submission was the most common. A significantly increased crude mortality rate was seen with ‘an admission to the ICU’ and ‘an ISS > 15’. Conclusion: We identified 17 RTRs, analysed the differences among RTRs and investigated the effect of the inclusion criteria on patient demographics. These findings may be useful when improving or developing RTRs. ß 2012 Elsevier Ltd. All rights reserved.

Keywords: Trauma registry AIS ISS Inclusion criteria Data collection Governmental funding

Introduction A trauma registry is a comprehensive data repository regarding injured patients. A trauma registry contains comprehensive information related to injury, which is generally unavailable in general hospital admission or emergency department data collections; therefore, a trauma registry aids not only injury research but also the development and evaluation of an injury prevention programme and the creation of clinical guidelines.1 Because a trauma registry is generally established in a hospital, it can also be used to monitor the hospital’s trauma care performance.1 Although a hospital trauma registry can evaluate the hospital’s trauma care performance, it is difficult to compare the performance

* Corresponding author at: School of Primary, Aboriginal and Rural Health Care, The University of Western Australia, M516 The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia. Tel.: +61 8 9346 2426; fax: +61 8 9346 1665. E-mail addresses: [email protected] (H. Tohira), [email protected] (I. Jacobs), [email protected] (D. Mountain), [email protected] (N. Gibson), [email protected] (A. Yeo). 0020–1383/$ – see front matter ß 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.injury.2012.08.024

of a given hospital with that of other hospitals. A regional trauma registry in which multiple hospitals participate could provide information regarding the relative performance of a given hospital.2 This type of trauma registry could include a state or province, nation and/or many nations. Another benefit of a regional trauma registry is that it may provide epidemiological data regarding injury in a wider geographical area than a hospital trauma registry. Such epidemiological data may be useful for policy makers who are creating injury prevention programmes by allowing them to determine a target cohort and monitor the effectiveness of the programmes.3,4 One of the most important factors of regional trauma registries is the funding source because the development and maintenance of a regional trauma registry has a considerable cost, and as the number of case registrations increases, so will the maintenance cost.5 The maintenance cost of a regional trauma registry was reported to be Australian $100 per case6; therefore, to sustain a regional trauma registry, a funding source should be secured. The version of the Abbreviated Injury Scale (AIS) used in the trauma registry is also an important factor because the version can affect the distribution of the injury severity scores.7,8 Most trauma registries use the AIS 90, AIS 98, AIS 2005 or AIS 2008. The AIS 90

H. Tohira et al. / Injury, Int. J. Care Injured 43 (2012) 1924–1930

and AIS 98 are compatible in terms of the value of the Injury Severity Score (ISS).9 The AIS 2005 and AIS 2008 are also compatible in most cases.10 However, ISSs based on the AIS 2005 or AIS 2008 were reported to be significantly less than those based on the AIS 98.7,8,11–13 The use of the AIS 2005 reduces the number of case registrations if an ISS > 15 is included in the inclusion criteria, compared to the use of the AIS 98.7,8,10 This reduction in case registrations might be apparent in patients with brain injuries because the AIS 2005 and AIS 2008 included many added or modified codes with a lower severity level than that of the AIS 98 for common brain injuries (i.e., subdural haematoma).10,11,14 The AIS versions used in regional trauma registries have not yet been investigated. The method of data registration needs to be considered when reforming an existing registry or developing a new registry. In contrast to a hospital-based trauma registry, a regional trauma registry requires a system that accurately and efficiently collects injury data from geographically separated hospitals. Three types of data collection were reported, including paper-based data collection, on-line data transfer and web-based data collection.15 Considering the rapid advancement of Internet technology, the data collection methods based on the Internet may be more common than ever before. The inclusion and exclusion criteria were reported to be inconsistent across regional trauma registries.1,15 Bergeron et al. demonstrated that different inclusion and exclusion criteria for trauma registries resulted in significant changes in the demographics of selected cohorts.16 For example, the use of an ISS < 12 as an exclusion criteria excluded 75% of the total trauma population, 65% of the total hospitalisation days and 41% of the total deaths. These exclusions from trauma registries may result in an underestimation of resource utilisation of trauma centres and represent only a limited view of injury epidemiology.17 Other criteria that may influence the demographics of registries are not fully understood. Understanding the differences in the structure, inclusion and exclusion criteria and demographics of regional trauma registries can provide ideas for future improvement and recommendations for those who are planning to develop new regional trauma registries. A few authors have conducted comparative studies of regional trauma registries.1,15 These studies revealed many differences among registries; however, they found that seven regional registries at most did not report the AIS versions that were used in the registries and did not investigate the effect of the inclusion criteria. Considering this situation, we formulated the following study questions: (1) What is the most common funding source for a regional trauma registry? (2) Which AIS version is used to code injuries in a regional trauma registry? (3) What is the most common data collection method for a regional trauma registry? (4) How do the inclusion and exclusion criteria affect the demographics of regional trauma registries? To address these questions, we aimed to identify existing regional trauma registries in the world and describe their structural differences, inclusion and exclusion criteria and demographics, as well as compare the demographics of registries with a certain inclusion criterion to those without the criterion. Materials and methods Case definition In this study, we considered a national trauma registry to be a regional registry. If there was no national registry in a given country, a state or provincial trauma registry was included. A regional trauma registry that covered multiple countries was also included.

1925

Search strategy We searched for articles associated with regional trauma registries using the MEDLINE database and a general Internet search engine (Google1) in July 2011. We used ‘trauma registry’ and ‘trauma database’ as search words (Fig. 1). We set no language restrictions and utilised a translation service if required for nonEnglish documents. Relevant articles were those that described a regional trauma registry and/or compared a regional trauma registry with others. We used the references of the relevant articles to identify other potentially relevant articles. We used a similar approach to identify homepages, annual reports and other relevant documents of regional trauma registries by using the general Internet search engine. Data abstraction Using all of the articles and other relevant documents, we abstracted the information regarding the structure, inclusion and exclusion criteria and demographics of registries. For the structural information, we collected the year of establishment, funding source, versions of the AIS that were used in a given registry, method of data submission and mode of participation. For the demographics, we extracted the total number of registered cases, number of case registrations per year, number of participating hospitals/institutions, proportion of patients aged 65 years or older, proportion of males, proportion of patients with an ISS > 15, proportion of penetrating injuries, leading cause of injury and crude mortality rates. We also computed the number of registered cases per hospital per year by dividing the number of case registrations per year by the number of participating hospitals. When these demographic data were not available, we used other data to compute the variables. For example, a crude mortality rate could be computed by dividing the number of deaths by the total number of cases. When the information that we sought was neither available in nor computable from collected data, we contacted the authors of the relevant articles or the data custodians of the registries. Comparison of demographics We compared the demographic data that included a certain criterion with those gained from the registries without the criterion to investigate how the inclusion criterion affected the patient demographics of the registries. We divided the registries into two groups according to the following three conditions: 1. Is ‘ISS > 15’ included in the inclusion criteria? 2. Is ‘admission to ICU’ included in the inclusion criteria? 3. Is ‘transferred patient for higher care’ included in the inclusion criteria?

Set 1 2 3 4 5 6 7 8

Search terms (trauma registr*).tw

(trauma database*).tw #1 or #2 comment.pt letter.pt Limit 1 to animal #4 or #5 or #6 #3 Not #15 Fig. 1. Search strategy to find relevant articles.

H. Tohira et al. / Injury, Int. J. Care Injured 43 (2012) 1924–1930

1926

These conditions were selected because they seemed to affect the demographics of the registries. Furthermore, the number of registries that included these conditions was similar to the number of registries that did not. We selected the following three outcome measures for comparison: the number of cases per hospital per year, crude mortality rate and proportion of patients with an ISS > 15. These three outcome measures were selected because they provided fundamental information regarding the injury data and were available in most registries. For each of the three conditions, we compared these outcome measures for the registries that fulfilled the condition with those for the registries that did not. Statistical analysis Because of the small sample size, Mann–Whitney tests were performed with the Predictive Analytics Software (PASW) Statistic18 version 18.0.3 (IBM SPSS, Inc., 2010, Chicago, IL). We set the significance level at 5%, and we excluded registries from the statistical comparisons if the information regarding the outcome measures was not available. Results Search results We identified seventeen regional trauma registries, which consisted of eleven national, four state and two multinational registries (Table 1). We could not identify a regional trauma registry in South America or Africa. Structures The structures of the identified regional trauma registries were shown in Table 2. The most common funding source was governmental funding (8 registries). The Trauma Audit & Research Network (TARN) of the UK was the only self-funded registry. The AIS 98 and AIS 2008 were most commonly used (four registries

each), followed by the AIS 90 (three registries) and AIS 2005 (two registries). The National Trauma Registry (NTR) of Canada currently uses the AIS 90 but will start to use the AIS 2008 in 2012. A web-based data submission was the most common data submission method. Inclusion and exclusion criteria The inclusion and exclusion criteria of the regional registries were listed in Table 3. An ISS > 15, admission to the ICU and transferred patient were the most common inclusion criteria (six registries each). Although hospitalisation was also commonly found in the inclusion criteria, hospitalisation was further divided into subcategories according to the length of stay (LOS) (e.g., LOS > 24 h, LOS > 3 days). A few registries excluded dead-onarrivals, deaths at the scene and isolated femoral neck fractures. Ten registries did not set exclusion criteria. Demographics The demographical data of each regional registry were presented in Table 4. The number of case registrations varied considerably. The National Trauma Data Bank (NTDB) in the U.S. exhibited the most case registrations (4 million) and the most case registrations per year (680,000 in 2009). The Israel trauma registry had the most case registrations per hospital per year (1316 in 2005), followed by the NTDB (997 in 2009). The proportion of patients aged 65 years or older ranged between 20 and 30% for the majority of registries. The proportion of patients with an ISS > 15 varied considerably, ranging from 8.6 to 86%. Penetrating injuries accounted for approximately 5% of registered cases for most registries, while the NTDB exhibited the highest proportion of penetrating injuries (9.5%). The male gender was dominant, and road traffic accidents were the leading cause of injury for the majority of registries. The National Trauma Database (NTrD) in Malaysia showed the highest crude mortality rate (31.2%), whereas the French trauma registry demonstrated the lowest crude mortality rate (1.0%).

Table 1 List of regional trauma registries. Country

Registry

Internet home page

Asia

Japan Malaysia UAE Israel

Japan Trauma Data Bank (JTDB) National Trauma Database (NTrD) UAE Nationa Trauma Registry Israel National Trauma Registry

http://www.jtcr-jatec.org/traumabank/index.htm http://www.acrm.org.my/ntrd/ Unknown http://www.gertnerinst.org.il/e/health_policy_e/ trauma/trauma_traumaregistry/

North America

US Canada

National Trauma Data Bank (NTDB) National Trauma Registry (NTR)

http://www.facs.org/trauma/ntdb/index.html http://www.cihi.ca/CIHI-ext-portal/internet/en/ document/types+of+care/specialized+services/ trauma+and+injuries/services_ntr

Europe

UK

United Kingdam Trauma Audit & Research Network (UKTarn) Trauma register DGU Greece national trauma registry French Trauma Registry Registro Intraospedaliero Multiregionale Traumi Gravi (RITG) Euro TARN

https://www.tarn.ac.uk/Login.aspx

Victoria state trauma registry (VSTS) New South Wales trauma registry (NSWTR)

http://www.health.vic.gov.au/trauma/registry.htm http://www.itim.nsw.gov.au/wiki/The_NSW_ Trauma_Registry_system http://www.uq.edu.au/conrod/index.html?page=100412 Unknown http://www.uq.edu.au/ntrc/

Germany Greece France Italy 14 countries in Europe Oceania

Australia

Australia and New Zealand

Queensland trauma registry (QTR) South Australian trauma registry (SATR) National Trauma Registry Consortium (NTRC)

UAE: United Arab Emirates, US: United States, UK: United Kingdom.

http://www.traumaregister.de/ http://www.trauma-society.gr/ http://www.rarr.inrets.fr/ http://www.cgsi.it/rit/homepage2.htm http://eurotarn.man.ac.uk/

H. Tohira et al. / Injury, Int. J. Care Injured 43 (2012) 1924–1930

1927

Table 2 Structures of regional trauma registries. Region

Country

Type

Establishment Funding

Asia

Japan Malaysia UAE Israel

National National National National

2004 2006 2003 1995

AIS

Participation Data items

Participating fee, Grant Web-based Gov Paper, web-based Unknown Web-based Unknown Web-based

AIS 98 Unknown Unknown Unknown

Voluntary Voluntary Voluntary Voluntary

92 36 100 300

National National

1993 1997

Gov, ACS, data fees Gov.

On-line data transfer On-line data transfer

Any upto AIS 2005 AIS 90 (–March 2012) AIS 2008 (April 2012–)

Voluntary Voluntary

107 46

UK

National

1989

Participating fee

250+

National

1993

Grants

AIS 98 (–May 2009) AIS 2005 (June 2009–) AIS 98

Voluntary

Germany

Voluntary

287

Greece France Italy

National National National

2005 1995 2004

Trauma society Unknown Gov.

Paper-based (1997–2005) Web-based (2005–) Web-based/paper/data transfer Unknown Paper Web-based

Voluntary Voluntary Voluntary

150 23 110–130

Euro TARN

Multinational 2002

International collaboration

On-line data transfer

Unknown AIS 90 AIS 98 (–2008) AIS 2008 (2009–) Unknown

Voluntary

237

Victoria

State

2001

Gov, data fee

Web-based

Voluntary

36

New South Wales

State

2002

Gov.

Paper (–2008) Web-based (2009-)

Voluntary

25

Queensland

State

1998

Gov, Insurance co., Univ. Gov, Uni. Societies and public sectors

Web-based

AIS 98 (–June 2010) AIS 2008 (July 2010–) AIS 98 (–June 2007) AIS 2005 (July 2007–2009) AIS 2008 (2009–) AIS 90 (1998–2008) AIS 2008 (2009–) AIS 98 Unknown

Voluntary

97

Voluntary Voluntary

95 13

North America US Canada

Europe

Oceania

1994 South Australia State Australia and Multinational 2003 New Zealand

Data submission

Paper-based Unknown

UAE: United Arab Emirates, US: United States, UK: United Kingdom, Gov.: governmental funding, ACS: American College of Surgeon, Uni.: University. AIS: Abbreviated Injury Scale. Table 3 Inclusion and exclusion criteria. Region

Country

Inclusion criteria

Exclusion criteria

Asia

Japan Malaysia

AIS  3 Deaths after injury ISS > 15 Admission to ICU with mechanical ventilation for 24 h Urgent surgery for head, chest, abdominal, spine and pelvis Severe head injuries (GCS 3–8) Deaths after hospital arrival Hospitalisation for >24 h All hospitalisations Deaths in ED Transferred patients

None None

UAEa Israel

North America

US

Injury diagnosis (ICD-9CM: 800–959.9) Hospitalisation Transferred patient Deaths after injury ISS > 12 and one of followings: Hospitalisation; Treatment in ED; or Died in ED.

None

UK

Hospitalisation for >3 days Transferred patient Admission to ICU Death within 30 days

Germany Greece France Italy

Admission to ICU Emergency life-saving surgery Hospitalisation DOA Transferred patient Victims of RTA All admission within 24 h after event; and ISS > 15 or early admission to ICU ISS > 15

Isolated femoral neck fracture (age > 65) Single pubic ramus fracture (age > 65) Single isolated injury None Injuries due to comorbidities None None

Canada

Europe

Euro TARN Oceania

None Deaths at scene DOA Non-hospitalisation Admission >72 h after injury event

Victoria

Deaths Admission to ICU with mechanical ventilation for 24 h serious injury to 2 or more body regions ISS > 15 Urgent surgery for head, chest, abdominal, spine and pelvis LOS > 3 days Transferred patients

New South Wales

ISS > 15 and admitted within 14 days of eventb Admission to ICU Deaths in hospital LOS > 24 h Transferred patients Deaths in ED Deaths during admission Vary with hospitals ISS > 15

Queensland South Australia Australia and New Zealand

None

None Isolated femoral neck fractures Isolated dislocation on upper limb, shoulder girdle, toe/foot/knee joint Isolated closed limb fracture Soft tissue injuries only Burns to <10% of BSA Isolated eyeball injury Isolated femoral neck fractures None Vary with hospitals DOA Deaths at scene

UAE: United Arab Emirates, US: United States, UK: United Kingdom. AIS: Abbreviated Injury Scale, ICD-9 CM: International Classification of Diseases, Ninth Revision, Clinical Modification. ISS: Injury Severity Score, ICU: Intensive Care Unit, GCS: Glasgow Coma Scale, ED: Emergency Department. DOA: dead on arrival, RTA: road traffic accident, LOS: length of stay, BSA: body surface area. a These criteria were defined for the Al-Ain Hospital in UAE. Criteria for the UAE national trauma registry were unknown. b An ISS > 12 is currently being used.

1928

Table 4 Demographics of regional trauma registries. Country

Case total

Cases per year

Hospitals

Cases per hospital per year

Age  65

Male

ISS > 15

Penetrating

Causes of injuries

Mortality

Asia

Japan

55,352 (2004–2009)

15,319 (2010)

147

102

39%

63%

39%

4.5%

Malaysia UAE Israel

Unknown 3500 (2003–2007) 200,000 (1997–2007)

1220 (2008) Unknown 25,000 (2005)

8 3 19

153 Unknown 1316

9.2% Unknown Unknown

88% Unknown 67%

79% Unknown 10%

5.9% Unknown 3.2%

Fall (40%, overall) RTA (44.9%, ISS > 15) RTA (74.4%) Unknown Fall (47%, overall) RTA (42%, ISS > 15)

10.9% (overall) 21.8% (ISS > 15) 31.2% Unknown 1.7% (overall) 12.2% (ISS > 15)

North America

US Canada

4 millions 111,297(1999–2008)

680,000 (2009) 14,065 (2008)

682 107

997 131

22% 28%

64% 71%

26% 86.3%

9.5% 5%

Fall (37%), RTA (30%) RTA (41%)

2.9% 11.0%

Europe

UK

23,840(2010)

110

217

29.8%

62.6%

36.4%

4.4%

Fall (44%)

Unknown

Germany Greece France Italy Euro TARN

10735 (2000–2003, ISS > 15) 51,425 8900 (2005–2006) 114,000 (1996–2008) 4397 20,582 (2000–2005)

9651 (2009) 8900 8342 (2008) 795 (2009) Unknown

218 32 160 3 14 groups

44 278 52 264 NA

Unknown 23% 7% (age > 60) 21% Unknown

70% 68.7% 63% 79% Unknown

65% 11.2% Unknown 72% NA

5% 0.47% NA 2.8% 4.0%

RTA RTA NA RTA RTA

12.7% Unknown 1.0% 17.5% Unknown

Victoria

40,000

2600 (2008)

138

19

23.5% (2005)

72% (2005)

84.5% (2008)

< 5%

New South Wales

18,000

2002 (2008)

14

143

27.9%

71.6%

NA

4.6%

Queensland

Unknown

16,691 (2009)

20

835

26.5%

63.8%

8.6%

6.5%

South Australia Australia and New Zealand

39,263 25,000 (2002–2005)

3206 6939 (2005)

6 4 states, 7 hospitals

534 NA

10.80% 22.1%

69.40% 72%

18% NA

5.6% 2.7%

RTA (23.6%, 2008) Fall (20.7%, 2008) RTA (38.5%) Fall (38.4%) Fall (44.1%, overall) RTA (48.3%, ISS  13) RTA RTA (48.8%)

Oceania

UAE: United Arab Emirates, US: United States, UK: United Kingdom. RTA: road traffic accident, ISS: Injury Severity Score, NA: not applicable.

(53.8%) (44%) (64%) (55%)

11.5% 13.1% 2.4% (overall) 20.7% (ISS > 15) Unknown 12%

H. Tohira et al. / Injury, Int. J. Care Injured 43 (2012) 1924–1930

Region

H. Tohira et al. / Injury, Int. J. Care Injured 43 (2012) 1924–1930

1929

Table 5 Comparisons of demographical data by three inclusion criteria. Inclusion criteria

Cases/hospital/year

Crude mortality rate

%ISS > 15

n

Median

p-Value

n

Median

p-Value

n

Median

p-Value

ISS > 15a

Yes No

4 8

148 244

0.28

5 6

13% 3%

0.02

3 7

79% 26%

0.02

Admission to ICU

Yes No

6 7

147 278

0.18

5 7

13% 3%

0.005

5 6

72% 19%

0.13

Transferred patient

Yes No

6 7

556 143

0.07

4 8

3% 12%

0.11

6 5

19% 72%

0.052

ISS: Injury Severity Score, ICU: Intensive Care Unit. The total number of registries changed among nine comparisons because some of registries were excluded due to missing information regarding demographical data in comparison. a The NTR was excluded from these comparisons because the NTR used an ISS > 12 as a criterion.

Comparisons of demographics The results of statistical comparisons were given in Table 5. We found that registries with ‘an ISS > 15’ and those with ‘an admission to ICU’ included in the inclusion criteria exhibited a significantly greater crude mortality rate than those that did not (p = 0.02 and 0.005, respectively). The criterion of ‘a transferred patient’ did not significantly affect any of the three selected measures at the 5% level; however, two p-values (p = 0.07 and 0.0052) might be small enough in this small sample size to suggest the association between the criterion and the two measures. We confirmed that registries with ‘an ISS > 15’ in the inclusion criteria had a significantly greater proportion of patients with an ISS > 15 than those that did not (p = 0.02). Discussion We identified seventeen regional trauma registries, which included eleven national, four state and two multinational registries. The most common funding source was the government, and the acceptable versions of the AIS for injury coding were different among the registries. The AIS 98 and AIS 2008 were the most common version. An ISS > 15, an admission to the ICU, a transfer for higher treatment and hospitalisation were most commonly found in the inclusion criteria. The leading causes, proportion of the elderly population, proportion of males and proportion of penetrating injuries were similar among the registries, whereas the number of case registrations and the proportion of patients with an ISS > 15 varied. We found that including ‘an ISS > 15’ and/or ‘an admission to the ICU’ in the inclusion criteria affected the crude mortality rate. We identified more regional trauma registries than previous studies.1,15 This increased identification may be due to our broad search method of using the MEDLINE database and a general Internet search engine. Despite our broad searches, we could not identify an existing regional trauma registry in South America or Africa. Although we did not set any language restrictions, we used only English search words. The use of English-only search words might have resulted in the failure to identify registries in those regions. There were several types of funding sources identified, and we found that half of the regional trauma registries were funded by the government. This tendency might be explained by the superiority of government funding in terms of continuity and predictability, which is required to sustain a trauma registry.18 Despite the superiority of government funding, the TARN and the Japan Trauma Data Bank (JTDB) collected contributions from participating hospitals without receiving government funds. The fees required by the TARN are determined by the annual new patient

attendance at each trust.19 The maximum fee is £8700 for a major trauma centre. Regarding the JTDB, all participating hospitals must pay annual membership fees (¥100,000) to the Japan Trauma Care and Research, which is an administrative organisation of the JTDB.20 Although both funding sources can sustain regional registries, government funding may be better when there are few participating hospitals in a regional registry because the fees can be expensive for the participating hospitals. We identified the following four AIS versions that were acceptable for injury coding: AIS 90, AIS 98, AIS 2005 and AIS 2008. The AIS 98 and AIS 2008 were the most common. We also found that AIS 98 codes in the trauma registries in Italy and Victoria, Australia, were converted to equivalent AIS 2008 codes by using a validated mapping Table,21 and the National Trauma Registry (NTR) in Canada planned to begin using AIS 2008 for injury coding in 2012.21 Considering the increased use of AIS 2008 codes, a regional registry should be developed or renovated so that the registry can accept AIS 2008 codes. Web-based data collection was the most common method used to submit data. Compared to paper-based data collection, webbased data collection can increase the accuracy of data by computerised error-checking functions, thereby reducing the chance of human error.22 The initial cost to develop a web-based data collection system can be expensive; however, after developing the system on a central data-collecting server, web-based data collection was found to yield a shorter case registration time, lower cost per case to maintain the system and lower rate of error occurrence than the paper-based data collection.22 Therefore, for those who intend to develop or update a regional trauma registry, web-based data submission may be a better option. A group in Europe is making an effort to establish standardised inclusion criteria for regional trauma registries.23 This group determined ‘a New ISS (NISS) > 15’ as the sole inclusion criterion. Our finding that the criterion of ‘an ISS > 15’ was significantly associated with the crude mortality rate might partly justify the use of ‘a NISS > 15’ as the sole criterion. However, we also found that the inclusion criterion of ‘an admission to the ICU’ was significantly associated with a higher crude mortality rate in the registry data. This finding suggests that ‘an admission to the ICU’ identified patients at risk of death who could not be identified by other criteria. Therefore, researchers and policy makers who are involved in running a trauma registry should consider whether ‘an NISS > 15’ is sufficient as a sole inclusion criterion and whether other criteria may also be needed. Limitations We were limited in our ability to collect information from the registries. When there was missing information, we contacted the

1930

H. Tohira et al. / Injury, Int. J. Care Injured 43 (2012) 1924–1930

authors of the relevant documents and/or the data custodians of the registries. However, we did not always receive responses from the authors or the custodians, resulting in missing structural and demographic data. Additionally, we were limited in searching for articles and/or documents not written in English. We had limited power to identify significant differences in the selected measures among the registries because of the small sample size. Conclusion We were able to identify seventeen regional trauma registries and describe the structure, inclusion and exclusion criteria and demographics of these registries. We found that the inclusion criteria were significantly associated with the patient demographics. These descriptions and findings may be useful when developing or renovating a regional trauma registry and when comparing data between registries. Conflicts of interest The authors have no commercial associations or sources of support that might pose a conflict of interest. Acknowledgements We would like to acknowledge those listed below for answering our inquiries: Ms Laura White, Ms Antoinette Edwards (UK), Dr Stefano Di Bartolomeo (Italy), Ms Brennan Maureen, Ms Patricia Sidhom (Canada), Ms Melanie Franklyn (Victoria), Ms Helen Thomas and Dr Ron Somers (South Australia). This article is based on a study first reported in the Journal of the Japanese Association for the Surgery of Trauma in Japanese.24 References 1. Moore L, Clark DE. The value of trauma registries. Injury 2008;39:686–95. 2. Nathens AB, Xiong W, Shafi S. Ranking of trauma center performance: the bare essentials. Journal of Trauma 2008;65:628–35. 3. Hosking J, Ameratunga S, Bullen C, Civil I, Ng A, Rodgers A, et al. Screening and intervention for alcohol problems among patients admitted following unintentional injury: a missed opportunity? New Zealand Medical Journal 2007;120:U2417. 4. Javouhey E, Guerin AC, Gadegbeku B, Chiron M, Floret D. Are restrained children under 15 years of age in cars as effectively protected as adults? Archives of Disease in Childhood 2006;91:304–8. 5. Shapiro M, Cole K, Keegan M, Prassad C, Thompson R. National survey of state trauma registries-1992. Journal of Trauma 1994;37:835–42.

6. Cameron PA, Finch CF, Gabbe BJ, Collins LJ, Smith KL, McNeil JJ, et al. Developing Australia’s first statewide trauma registry: what are the lessons. ANZ Journal of Surgery 2004;74:424–8. 7. Palmer CS, Niggemeyer LE, Charman D. Double coding and mapping using Abbreviated Injury Scale 1998 and 2005: identifying issues for trauma data. Injury 2010;41:948–54. 8. Salottolo K, Settell A, Uribe P, Akin S, Slone DS, O’Neal E, et al. The impact of the AIS 2005 revision on injury severity scores and clinical outcome measures. Injury 2009;40:999–1003. 9. Skaga NO, Eken T, Hestnes M, Jones JM, Steen PA. Scoring of anatomic injury after trauma: AIS 98 versus AIS 90-do the changes affect overall severity assessment. Injury 2007;38:84–90. 10. Palmer CS, Franklyn M. Assessment of the effects and limitations of the 1998 to 2008 Abbreviated Injury Scale map using a large population-based dataset. Scandinavian Journal of Trauma Resuscitation and Emergency Medicine 2011;19:1. 11. Barnes J, Hassan A, Cuerden R, Cookson R, Kohlhofer J. Comparison of injury severity between AIS 2005 and AIS 1990 in a large injury database. Annals of Advances in Automotive Medicine 2009;53:83–9. 12. Tohira H, Jacobs I, Matsuoka T, Ishikawa K. Impact of the version of the abbreviated injury scale on injury severity characterization and quality assessment of trauma care. Journal of Trauma 2011;71:56–62. 13. Tohira H, Jacobs I, Mountain D, Gibsons N, Yeo A. Comparisons of the outcome prediction performance of injury severity scoring tools using the Abbreviated Injury Scale 90 update 98 (AIS 98) and 2005 update 2008 (AIS 2008). Annals of Advances in Automotive Medicine 2011;55:255–65. 14. Carroll CP, Cochran JA, Price JP, Guse CE, Wang MC, Carroll CP, et al. The AIS2005 revision in severe traumatic brain injury: mission accomplished or problems for future research. Annals of Advances in Automotive Medicine 2010;54:233–8. 15. Probst C, Paffrath T, Krettek C, Pape H. Comparative update on documentation of trauma in seven national registries. European Journal of Trauma 2006;32: 357–64. 16. Bergeron E, Lavoie A, Moore L, Bamvita JM, Ratte S, Clas D. Paying the price of excluding patients from a trauma registry. Journal of Trauma 2006;60:300–4. 17. Brotman S, McMinn D, Copes W, Rhodes M, Leonard D, Konvolinka CW. Should survivors with an injury severity score less than 10 be entered in a statewide trauma registry? Journal of Trauma 1991;31:1233–9. 18. Gronbjerg KA. How nonprofit human service organizations manage their funding sources: key findings and policy implications. Nonprofit Management and Leadership 1991;2:159–75. 19. The TRAUMA Audit & Research NETWORK. Participant contribution 2011/2012. https://www.tarn.ac.uk/Content.aspx?c=51. Accessed on 9 March 2012. 20. Japan Trauma Data Bank. General Provisions and detailed regulations. http:// www.jtcr-jatec.org/traumabank/dataroom/data/jtdb_kisokusaisoku2010.pdf. Accessed on 9 March 2012. 21. Palmer C, Franklyn M, Read-Allsopp C, McLellan S, Niggemeyer L. Development and validation of a complementary map to enhance the existing 1998 to 2008 Abbreviated Injury Scale map. Scandinavian Journal of Trauma Resuscitation and Emergency Medicine 2011;19:29. 22. Weber BA, Yarandi H, Rowe MA, Weber JP. A comparison study: paper-based versus web-based data collection and management. Applied Nursing Research 2005;18:182–5. 23. Ringdal K, Coats T, Lefering R, Di Bartolomeo S, Steen P, Rise O, et al. The Utstein template for uniform reporting of data following major trauma: a joint revision by SCANTEM, TARN, DGU-TR and RITG. Scandinavian Journal of Trauma Resuscitation and Emergency Medicine 2008;16:7. 24. Tohira H. International comparison of regional trauma registries. Journal of the Japanese Association for the Surgery of Trauma 2012;26:28–35.