Article Text
Abstract
Relevant local injury epidemiology In Hong Kong, there were, on average, about 19 596 traffic crashes involving 157 deaths and 21 106 injured persons each year between 2006 and 2011. Scientific analyses were conducted by geographers and engineers primarily using the police crash database. Medical professionals have been analysing road traffic injury data from hospital discharge summaries. Moreover, community leaders have been trying to promote local safe communities.
Best practices This paper describes the effort of a multidisciplinary team to address road safety problems and to sustain road safety benefits through a public health approach. The multidisciplinary team comprised a geographer, an engineer, medical professionals and community leaders. The project covered four tasks, namely data integration, identification of hazardous road locations, crash analysis and engineering study, and knowledge exchange through various activities involving a WHO-designated local safe community.
Implementation The crash and hospital databases for a district in Hong Kong with 500 000 population were integrated. Based on the integrated database, the public health and people-based approach was adopted to identify hazardous road locations—hot zones—using geographical information systems. Specific hot zones having strong patterns of common factors were considered as treatable locations with a combination of low-cost remedial measures. The benefits of the project are sustained through various activities engaging the general public and major stakeholders.
Research agenda More research should be conducted on how institutional support, scientific research and community involvement can be fruitfully combined to achieve the ultimate goal of sustained road safety benefits for people at the community level.
Statistics from Altmetric.com
Introduction
This programme report aims to explain the onset of a multidisciplinary team as an example of best practices in road safety, and to give details of what it includes. In this section, some background for understanding traffic injury prevention in Hong Kong is provided. In the section ‘The project details’, details of the best practice about data integration, identification of hazardous road locations, crash analysis and engineering study, and knowledge exchange are presented. These four tasks are collectively described as the project. Results of the multidisciplinary project are given in the section ‘The implementation’, before a research agenda is presented in the final section.
Data are essential to any scientific research, which in turn will have value to policy makers and the general public. In Hong Kong, there were, on average, about 19 596 crashes involving 157 deaths and 21 106 persons injured each year on its 2 073 km of motorable roads between 2006 and 2011. At the end of 2011, Hong Kong had a population of 7.1 million;1 hence, the deaths and persons injured per 100 000 inhabitants were roughly about 2.21 and 297.11, respectively. The crash database of the traffic information system (TIS) with georeference has been available in Hong Kong for a long time. However, crash locations were not systematically geovalidated.2 In Hong Kong, any road traffic crash should be reported to the Hong Kong Police Force within 6 months of the incident under the Traffic Accident Victims Ordinance. All traffic crashes reported to the Police are stored in the Traffic Accident Data System (TRADS) data files, which in turn form part of the TIS. Demographic information, injury severity, the time of crash, conditions of roads, time and characteristics of crash of every incident were recorded. Injury severity is based on a simple classification of whether the traffic crash victims died within 30 days and as a result of the traffic injuries, or admitted to a hospital for more than 12 h (serious injury), or not (slight injury).
Health information can be gathered from the existing computerised data system and the new information system of injury surveillance from emergency department visits. Hospital discharge summaries, also called electronic patient records (ePR), store data from all hospital admissions and detailed patient conditions, procedures undergone and need for intensive care during hospital stay. Since it is used for medical purpose, environmental factors, such as weather, road and traffic conditions, are usually not documented. In Princess Margaret Hospital (PMH), the trauma centre in Kowloon West cluster serving a population of 0.5 million, all documented injuries are also registered under the new injury surveillance system (ISS). PMH is a major acute general hospital with 1172 beds and is equipped with two trauma resuscitation rooms in the accident and emergency (A&E) department. All major traumas happening within its service area are diverted to PMH with the trauma activation system in effect 24 h a day for all suspected severely injured patients according to predefined criteria. The ISS collects information on all visits to the PMH A&E department due to all sorts of injuries. As the ISS was at a pilot stage, it was not yet implemented in the other four trauma centres, which serve different geographical areas in Hong Kong. The data collected through the ISS includes type and location of the crash, demographic information of each patient, the mode of arrival, the triage category, diagnosis, procedures performed and outcome (discharged, discharged with referral to other specialties, admitted, and died).
Working primarily on the crash database, many geographers conducted road safety research by using their spatial analytical skills to identify and analyse hazardous road locations;3–5 and engineers are often concerned with improving road safety through the introduction of engineering measures and to evaluate their effectiveness.6 ,7 However, with the limited health information, scientific spatial analysis to identify hazardous road locations commonly take traffic crashes (events), rather than persons killed or injured in traffic crashes (people), as the primary object of analysis. Working primarily with the hospital database, many medical professionals worldwide focus on analysing injury patterns of road traffic casualties.8–10 Hence, diagnosis and treatment are given much attention. In particular, medical professionals often focus on saving lives and reducing severity of injury after crashes, and geographers and engineers often focus on avoidance of crashes. Nonetheless, it is increasingly recognised that better public health requires both.
Besides, the local community has to be involved to ensure that the message of dangerous road locations can be effectively disseminated to individual road users. The overall awareness of road safety also has to be raised. These are also the key aims of WHO's safe community efforts in injury prevention.11 ,12 The published indicators for safe transport stated the importance of partnership and collaborations by a group of managers, workers, technical staff, voluntary organisations and safety professionals who are responsible for promotion of safety in the traffic areas.11 ,12 In our study area, there is a WHO-designated Kwai Tsing Safe Community. However, there is a separation of scientific research, evidence-based interventions and community engagement. This project represents a meaningful effort of a local academic, an engineering practitioner, medical professionals and community leaders in Hong Kong to integrate the two existing databases, conduct scientific evidence-based research, recommend specific road safety improvement measures, and sustain the road safety benefits through the community-based local safe communities as advocated by WHO.
Project details
The multidisciplinary team
Multidisciplinary research and collaboration are highly important for the success of a public health approach on road safety strategy.11–15 A public health approach on road safety lays emphasis on raising awareness of road safety as a global health burden, identifying vulnerable groups, preventing traffic injury and saving lives.11 ,12 While the approach is not novel to medical professionals concerned about injury prevention, it was not common in geographical and engineering research. This project adopts the public health approach to road safety across disciplines and throughout its four major tasks of data integration, identification of hazardous road locations, crash analysis and engineering study, and exchange of knowledge. The multidisciplinary team consists of a geographer, an engineer, medical professionals and local community leaders. The geographer has spatial analytical skills and has developed the GIS-based hot-zone methodology based on the locational characteristics of traffic crashes. The engineer has expertise in road safety design, audit, inspection and crash analysis. Medical personnel from the hospital were heavily involved throughout the process to advise on injury patterns. They included a trauma nurse consultant, a medical consultant, an A&E specialist and the A&E chief of service. Moreover, a hospital administrator and a leader of the local community in WHO-initiated safe community initiatives were involved.
Integrating place-based with people-based safety analysis
Traditional analysis examines crashes as the primary objects; and the crash scenes were the locations commanding attention of geographers, engineers, the public health personnel, police and other emergency personnel like Fire Services. With the integration of the two databases in this study, we are able to treat and analyse people—the traffic crash victims—directly as the primary objects. In other words, the locations and clustering patterns of people killed or injured in traffic crashes can be analysed directly. During the study period of 2003–2007, there were 4364 crashes and 5912 persons injured or killed within our study area. Its significance, as reflected by the shares of traffic crashes (5.84% of 74 688) and persons injured or killed (6.20% of 95 430), was roughly in proportion with its significance in population (7.04% of 7.1 million). Several types of injury-inflicting traffic crashes having particularly high implications from a public health perspective are identified for detailed analysis. The first one was multiple-casualties crashes. Though a multiple-casualties crash may appear as one point only in a map with the traditional analysis, many people were hurt. More importantly, these bigger incidents are likely to trigger the resuscitation team and require triage in cases where the nearest hospital does not have enough capacity to deal with the large numbers of patients requiring emergency treatment. The second one was pedestrian crashes. Earlier analysis demonstrated that pedestrians, particularly the elderly hurt in traffic crashes, would have a much higher possibility of severe injury or even fatality.16 In Hong Kong, the ratio was 2.77 times (ISS>15), 5.16 times (ISS>30) and 3.61 times (mortality) for those aged 65 years and above, as compared with those aged 15–64 years old.16 The last one was serious or fatal crashes. These traffic casualties are in urgent need of medical treatment, and the outcomes may depend greatly on the time of arrival to the crash scene and the provision of timely treatment in the first 3 h. Perceiving traffic injuries as a public health problem, the crash-based hot-zone methodology has been modified based on traffic casualties (people) rather than traffic crashes (event). These are called casualty-based hot zones.
Crash analysis and engineering measures
For each hot zone, the crash and hospital data of every crash and person killed or injured were analysed with respect to the characteristics of road, traffic, injury pattern and other relevant factors, such as crash type, weather and other circumstances as recorded in the TRADS database. In the crash analysis, logistic regressions were conducted to obtain the OR of different selected explanatory variables from TRADS and ISS with the outcome variable as ‘serious or fatal injury’ or ‘slight injury’. The main objective of analysing the integrated database for engineering measures is to establish the presence of any patterns and to define any particular safety problem at the site. In conjunction with crash analysis, site inspection was conducted to identify additional relevant factors in road design, traffic conditions, signs and markings, as well as road users’ behaviour (such as jaywalking), through direct observation at the site.
Upon confirmation of a site-specific problem, an array of engineering solutions from road safety audit guidelines and engineering manuals were screened for possible treatment.17–21 Remedial engineering measures are very much site-specific. There is a tendency to over-provide treatments which could degrade the streetscape and even generate new safety problems. Moreover, it is recognised that hot zones may arise due to statistical randomness; and treatments are subject to the regression-to-mean effect which ‘overstate’ the evaluation results.5 ,22 For this reason, the optimal solutions were recommended by following a series of steps from analysing patterns, identifying common factors to determining treatability.
Sustained efforts through hospitals and the safe community efforts
Results have to be shared with decision makers in charge of collecting crash data, providing emergency services, providing medical treatments and improving road safety. The project fulfils the requirement of establishing an operational programme to document the frequency and causes of traffic injuries. The efforts are based on the intersectoral approach to raise awareness of the safety problem. Sharing knowledge and raising awareness in this project were only the first steps. It is suggested that further resources need to be applied to develop, introduce and evaluate evidence-based preventive actions to improve road safety in the specific local context.
Implementation
The integrated database
The major steps of this study in integrating TRADS with ISS are described below. TRADS has the Accident Report Booklet (ARB) number and both the ISS and ePR have the A&E number, which is assigned to each patient visiting the A&E departments. Based on the ISS, each road traffic injury patient is matched with a stepwise strategy for linkage by demographic data to the ARB number and, subsequently, information from discharge summaries are searched through ePR using the A&E number. The A&E number is then also used as a record linkage key. These data include road types, objects of impacts and occurrence of rollover. The precise location of each crash is determined by using a six-step GIS-based validation process;4 and newspaper archives from university libraries in Hong Kong were used to ascertain the locations as the last step in manual checking. For injury data, injury severity takes the following forms: abbreviated injury severity (AIS) range, maximum AIS, ISS and number of injuries.
Results of casualty-based hot zones
The hot-zone methodology identifies hazardous road locations consisting of more than one contiguous road segment defined by basic spatial units (BSU).4 ,23 ,24 For events (such as crashes) happening on road networks, a BSU (ie, a road segment) represents the unit of analysis for calculating event intensities and identifying spatial clusters. In road safety research, ‘blocks’ of road segments with higher crash intensities are called hot zones. The methodology emphasises the spatial contiguity of the road network in the identification of hazardous road locations. It recognises that road segments are not spatially independent objects, and the length of hot zones is determined by the actual clustering tendency of traffic crashes rather than fixed a priori. In addition, the determination of hot zones may be adjusted by weighting the number and severity of injuries, or pooled data, over multiple years.
To recall, casualty-based hot zones in this study are classified by three defining criteria: killed or severely injured (KSI), multiple casualties and pedestrian casualties. Each hot zone refers to an aggregate of densely populated injury cases with two or more BSUs. The standard length of a BSU is 100 m.4 Hot zones were derived from registered number of persons killed or injured from the TRADS system between 2003 and 2007. A KSI hot zone is defined with a threshold of six persons killed or seriously injured as registered in the TRADS system. A multiple-casualties hot zone is defined as 28 or more persons injured. A pedestrian hot zone is defined as eight or more pedestrians injured. In this study, numerical, instead of statistical and model-based definitions of the threshold values are used for illustration purposes, and in view of resources required for the engineering site visits.20 The distribution of the identified hot zones is shown in figure 1. It can be seen that some of the KSI hot zones were also multiple-casualties and pedestrian hot zones.
Of all traffic casualties injured in the identified KSI hot zones (n=223), pedestrians (n=34) were associated with higher likelihood of KSI according to the TRADS system (OR=2.15, 95% CI 1.8 to 2.5). Across the KSI hot zone sites, 88% of pedestrian casualties were classified as serious, while the remaining 12% were classified as fatal. Among traffic crashes involving automobiles (n=172), the risk of drivers sustaining a serious or fatal injury is about twice that of their passenger counterparts (OR=2.05, 95% CI 1.5 to 2.8).
Totally, 10 multiple-casualties hot zones with 28 or more casualties from 2003 to 2007 were identified (figure 1). Among these multiple-casualties hot zones, 23% of pedestrian casualties were classified as seriously injured or fatal, while the remaining 77% were classified as having slight injuries. Among traffic crashes involving private vehicles (n=373), the risk of rear seat passengers sustaining a slight injury was slightly higher than that of their front seat counterparts (OR=1.1, 95% CI 1 to 1.2).
Finally, 13 pedestrian hot zones with eight or more pedestrian casualties in the 5-year period were identified (figure 1). Of all casualties identified in the pedestrian hot zones (n=483), pedestrians (n=75) were associated with a higher likelihood of KSI according to the TRADS system (OR=1.6, 95% CI 1 to 2.6). The observed likelihood, however, was lower than the comparable odds observed among injuries which took place at the KSI hot zones.
Engineering measures
Site visits were conducted at each site to record the traffic engineering features and traffic operating conditions. Photographs were taken for subsequent analysis. For highways not accessible on foot, video movies were taken. The main traffic engineering features of interest are listed in table 1. Based on table 1, an assessment was conducted on the 24 hot zones (depending on the actual crash patterns, a hot zone may belong to more than one type of the KSI, multiple-casualties and pedestrian hot zones). Two of them were excluded from further studies because many crashes actually happened within logistics buildings, as the Hong Kong container port is located within the study area.
For hot zones, it is a priority to address those sites which have a clear pattern of crashes. A clear pattern may be a repeated happening of the same crash type (such as rear-front crashes) or casualty role (such as pedestrians). Four categories of patterns from no (0), weak (+), moderate (++) to strong (+++) patterns were used. Depending on the crash type, the key road traffic engineering features were examined, and key common factors, such as poor visibility, downhill gradient, lack of space for pedestrians, were identified. Given the existence of a pattern and common factors, treatability of the problem is another issue. In considering treatability at the sites, a range of low-cost engineering solutions were considered. These low-cost measures are defined as measures which do not require extensive engineering works, such as continuous road widening, realignment or major construction/alteration of structures.17–19 They generally only involve the use of signs, markings or signals, minor adjustment or modification of street furniture or local widening or adjustments to road layout. Treatability is evaluated depending upon whether the site is conducive for treatment with low-cost measures which are viable, proportionate to the problem and effective.17–19 Each hot zone was given a treatability score from low (1) to high (3). A score of 1 means that the site is difficult for administering treatment using conventional measures; a score of 2 means that the site is treatment-friendly using more extensive/drastic measures; and a score of 3 means that the site is amenable to treatment using low-cost, fast-track measures. Admittedly some safety problems are easier and far less costly to treat than the others. In some situations, there was substantial uncertainty about the appropriate solution, and a more detailed study was recommended.20 ,21 Expert opinions had to be used in conjunction with the road safety guidelines.20 The engineer in this team has undergone road safety audit training in Europe; and he is an experienced and practising road safety engineering consultant.
After the hot zone assessment, recommendations for improvement fall into two categories: specific recommendations for each hot zone, and general recommendations. Specific recommendations are given for each hot zone with respect to the crash data and the specific traffic engineering and operational characteristics identified in site inspections. It is noteworthy to mention that many of these recommendations are low-cost solutions which can be applied in the short term. However, medium- to long-term solutions are also given if these are considered appropriate. General recommendations are formulated over a broader picture of road traffic casualties in Kwai Tsing and, where appropriate, the territory as a whole. They are generally more applicable in the medium to long term.
Preliminary efforts to sustained partnership in road safety
Throughout this project, regular discussion and sharing of findings were made by PMH with members of the District Council and WHO-designated Kwai Tsing Safe Community. In relation, the Traffic Injury Prevention Symposium 2010 was held. The event was well attended with more than 150 participants, many of whom were medical professionals working with the hospital authorities. Moreover, through the sharing of best practice by the Kwai Tsing Safe Community, the Sham Shui Po Safe Community, and the Tseung Kwan O Safe Community (two other WHO-designated safe communities in Hong Kong), have recognised the importance of injury surveillance, and traffic injury is part of their focus.
In addition, project reports were distributed to the District Council and Transport Department of the Government. Through sharing the results with key decision makers in charge of collecting crash data, providing emergency services, giving medical treatments, and improving road safety, this project has began to obtain some institutional support and recognition. A working group has been formed with members from academics, hospital authorities and the Transport Department to improve traffic injury surveillance in Hong Kong. The main focus is for road safety promotion. In particular, the Transport Department has taken the lead in fostering cooperation with academics and the hospital authorities for the feasibility of setting up a territory-wide database system for road traffic injuries. A feasibility study will be conducted in the coming months. The project's findings, especially on pedestrian safety, were further disseminated in a press release for the global launch of the United Nations Decade of Action for Road Safety in May 2011.
Research agenda
A major barrier to fostering an integrated public health approach to address traffic injury challenges is the separation of scientific research, medical expertise, engineering investigations and community engagement. This programme report is a systematic presentation of how multidisciplinary scientific research and community involvement can be fruitfully combined to achieve the ultimate goal of sustained road safety benefits for people at the community level. However, more research is still required. In the first place, accurate information about crashes, and their subsequent health consequences, is essential for monitoring road safety problems. Methodological problems, such as poor database design and under-reporting, and/or the lack of linkage between crash and public health databases remain key areas requiring worldwide attention.2 ,10 ,25–29 It is recommended that different institutions responsible for collecting traffic injury data (such as the police, transport authorities, hospitals, clinics and insurance companies) should be in direct contact with each other through an institutional setup with regular working meetings to advise policy makers on traffic injury matters. Better use of the integrated database through utilising more police- (such as speeding) and hospital-based (such as blood alcohol content) information for hot zone identification represents an important area of further research. Second, WHO's community-based approach to public health in engaging the general public is highly valuable.11 ,12 More efforts should be made to develop, implement and evaluate evidence-based preventive actions. At the community level, this project made efforts to conduct exchange of knowledge rather than to evaluate specific measures of intervention. The latter is important to any recommendation of evidence-based preventive actions. A platform for sustained partnership among different institutions responsible for collecting traffic injury data, the scientific community, and all relevant public and private non-governmental organisations should be fostered. In the long term, this platform can facilitate the development of a road safety strategy which represents a high level of governmental commitment to tackle the burden of traffic injury effectively with the active participation of different stakeholders.13–15
Key points
-
A multidisciplinary team worked together to integrate the crash and hospital databases for a district in Hong Kong with a population of 500 000.
-
A public health and people-based approach was adopted to identify hazardous road locations using geographical information systems.
-
Hazardous road locations having strong patterns of common factors were considered treatable with a combination of low-cost remedial measures.
-
The benefits are sustained through exchange of knowledge involving a WHO-designated local safe community.
New president should reform gun legislation
Mat Goldstein, writing in the Daily Trojan, noted that because nearly 30 000 people die from gun-related injuries each year in the USA, ‘one would think that gun violence and regulation would be important election issues.’ But neither Obama nor Romney saw fit to focus on gun control. It is now possible for firearms to be bought at gun shows without a background check, and a ban on assault weapons expired 8 years ago, leaving semi-automatic firearms available for unrestricted purchase. The writer acknowledges, ‘Not every firearm-related tragedy would have been stopped with stricter regulations in place but … many would have.’ And polls in 2007 and 2008 reveal that 87% of Americans support background checks on private sales of guns and 82% support limiting the sale of military-style assault weapons.
Editor: With this level of public support it seems the time has come to challenge the gun lobby and introduce long overdue regulations.
Quebec workplaces progressively safer
In November, two workplace explosions in Quebec killed several employees. These were not isolated events but they prompt us to look at the long-term data. Some statistics suggest that work-related fatalities in Quebec have greatly decreased in the last few years and that injuries have fallen by 36%, giving Quebec the fourth lowest rate in Canada. It has been observed that most events occur in small and medium size companies, where prevention is often ignored. However, one critic argues that ‘Quebec's system is rigged to minimise the number of (events) reported and to systematically favour employers in the claims process.’ The critic adds, ‘We hear a lot about awareness campaigns but (too) often these consist of slapping a poster in the cafeteria reminding people to wear their goggles.’
fshalom@montrealgazette.com
Acknowledgments
We are grateful for all the support and positive comments received throughout this project. Moreover, this project has won the Outstanding Community Partnership Award 2011.
References
Footnotes
-
Contributors All authors have contributed to the conception and design of this study. Acquisition of data, analysis and interpretation of data were mainly conducted by Loo, Leung, Kwong and Chow. Research dissemination efforts were mainly made by Chow, Lai and Chau. Drafting of the article and revising it critically were mainly done by Loo, Kwong and Leung. Final approval was obtained from all authors. All authors take joint responsibility for the integrity of the work as a whole, from inception to published article.
-
Funding This work was supported by the Hong Kong SAR Government Health Care and Promotion Fund grant number (21070504) and the Research Grant Council General Research Fund, grant number HKU 749210 H.
-
Competing interests None.
-
Patient consent Obtained.
-
Provenance and peer review Not commissioned; externally peer reviewed.
-
Data sharing The data used in this study were from the Transport Department and Hospital Authority of the Hong Kong SAR Government. Interested readers may write to these authorities for the data. However, the authors are not authorised to redistribute them.