Safety Science 59 (2013) 135–140 Contents lists available at SciVerse ScienceDirect Safety Science journal homepage: www.elsevier.com/locate/ssci Cyclists and open vehicle doors: Crash characteristics and risk factors Marilyn Johnson a,b,⇑, Stuart Newstead c, Jennie Oxley c, Judith Charlton c a Institute of Transport Studies, Department of Civil Engineering, Monash University, Australia Amy Gillett Foundation, Australia c Monash University Accident Research Centre, Monash Injury Research Institute, Australia b a r t i c l e i n f o Article history: Received 22 June 2012 Received in revised form 15 April 2013 Accepted 22 April 2013 Keywords: Cyclist safety Car door crashes Cyclist–driver interaction Naturalistic cycling study a b s t r a c t A cyclist crash with an open vehicle door can result in serious injury, sometimes fatal, outcomes. However, little is known about the frequency and range of injury outcomes of this crash type or the contributing factors. In this study, the factors associated with cyclist-open vehicle door collisions in Victoria, Australia were investigated using three complementary data sources: (1) police reported crashes (2000–2011); (2) hospital presentations (2000–2010), and (3) a sample of video footage from a naturalistic study of commuter cyclists in Melbourne (2009–2010). A total of 1247 police reported cyclist crashes and 401 hospital emergency department presentations were analysed. As a proportion of all cyclist crashes, cyclist-open vehicle door crashes accounted for 3.1% (hospital) and 8.4% (police). The majority of cyclists injured were: male (police: 67.1%; hospital: 65.8%); adults aged 18 years or older (police: 97.5%; hospital: 96.3%), and crashes occurred in speed zones up to 60 kph (police: 93.1%). From the naturalistic cycling study, there were 13 door-related events with a rate of 0.59 events per trip. No collisions occurred and in all cases, a potential collision was avoided by the cyclists’ evasive action. Most drivers did not look in the direction of the cyclist before opening the door. While the number of reported crashes is relatively low, cyclists’ exposure to events with potential for this crash type is high. Potential countermeasures to reduce the risk of this interaction/crash type are discussed and include road environment design improvements and road user behaviour programs. Ó 2013 Elsevier Ltd. All rights reserved. 1. Introduction Collisions involving cyclists impact with an open vehicle door can result in significant injuries, however, little is known about the nature and extent of these collisions, or injury outcomes. Cyclist-open vehicle door crashes are usually referred to as a subgroup of collision type in a list of crash mechanisms (Haileyesus et al., 2007; Wood et al., 2009). Specifically, this collision type, also referred to as being ‘doored’, is when the door of a vehicle that is parallel parked on the side of the road is opened into the path of an approaching cyclist creating the potential for collision (Dennerlein and Meeker, 2002). Cyclist crashes with unexpectedly opened vehicles doors can have fatal outcomes. Since 1989, in Victoria, four cyclists have died following a crash with a vehicle door (VicRoads, 2012). However, while this crash type may appear infrequent, like all cyclist crash types, it is difficult to comprehensively determine the total number of cyclists as cyclist crashes, in general, is grossly under-reported, ⇑ Corresponding author. Address: Institute of Transport Studies, Department of Civil Engineering, Monash University, Wellington Road, Clayton, Vic 3168, Australia. Tel.: +61 3 9902 0353; fax: +61 3 9905 4363. E-mail addresses: Marilyn.johnson@monash.edu.au (M. Johnson), Stuart.newstead@monash.edu.au (S. Newstead), Jennie.oxley@monash.edu.au (J. Oxley), Judith.charlton@monash.edu.au (J. Charlton). 0925-7535/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ssci.2013.04.010 especially when there is little or no vehicle damage (Atkinson and Hurst, 1983; Harris, 1990; Pucher and Dijkstra, 2003; Harman, 2007). Internationally, the reported rate of cyclist-open vehicle door collisions has varied. In New Zealand in the 1970s only 1 per cent of cyclists were vehicle door-related, however, there was little community awareness that this crash type was reportable and this was likely to have contributed to the low reported rate (Atkinson and Hurst, 1983). More recent studies have reported higher proportions of this collision type. In the 1990s in Toronto, Canada, being doored was the third most frequent cyclist crash type (11.9%) and most crashes involved adult cyclists (aged 20– 40 years) (Tomlinson, 2000). Among bicycle messengers in Boston, United States, being doored accounted for 16 per cent of all self-reported crashes (Dennerlein and Meeker, 2002). Research on bicycle messengers in the UK and Europe reported that to avoid being doored, bicycle messengers would engage in illegal behaviour, including riding into oncoming traffic (Fincham, 2006). In 2004, the City of Melbourne in Victoria, Australia identified that unexpectedly opened vehicle doors was a crash risk to cyclists, specifically in relation to bike lanes (City of Melbourne, 2004). In a separate report by VicRoads, the state roads authority, it was acknowledged that the ‘width of the [bike] lane is a compromise’ for cyclist safety (Daff and Barton, 2005:4). While it might seem 136 M. Johnson et al. / Safety Science 59 (2013) 135–140 reasonable to simply widen bike lanes to afford cyclists more space to avoid opening car doors, the authors suggested that this may result in drivers using the bike lane as a proxy vehicle lane and this would compromise cyclist safety. Anecdotally, this crash type is increasing, however there has been little targeted analysis to determine the rate of this particular cyclist crash type. Currently, there is no national database in Australia that can be queried to determine the frequency of cyclist crash types. At a state level, in Victoria, there are databases that report the police reported crashes (VicRoads CrashStats) and hospital presentations (Victorian Injury Surveillance Unit). Both of these data sources were analysed in this study. The aim of this study was to identify the extent and type of cyclist injury outcomes sustained in a cyclist-opened vehicle door crashes and to investigate factors associated with these crashes. The implications of the findings are considered within the Safe System Framework (SSF), the theoretical model that underpins road safety management in Australia. study, participants had a compact video camera attached to their helmets and each participant recorded 12 h of footage of their trips to and from work over a 4 week period. Cyclists rode their typical route with variations during the study period (Johnson et al., 2010). For the current analysis, a subset of footage was reviewed for 11 participants (eight males, three females). Two trips from each participant, one morning commute and one afternoon commute, were randomly selected and analysed, focusing on the interaction between the cyclist and opened doors of vehicles that were parallel parked. The behaviour of the driver and cyclist before, during and after the event, the road environment and the surrounding traffic were examined. The speed of the cyclist may be a contributing factor, however cyclist speed was not recorded as part of this study. In addition, it was not possible to accurately analyse the number of vehicles that were occupied. While an occupied vehicle is clearly a greater risk to a cyclist than an empty vehicle with no attendant road user, vehicle occupants were not visible from the recorded video footage. 2. Methods This study analysed three complementary data sources: police reported crashes, hospital emergency department presentations and naturalistic cycling video footage to investigate cyclists’ collisions, near collisions and injury outcomes involving open vehicle doors. 2.1. Police reported crash data Police reported cyclist crashes were analysed for the period from January 2000 to December 2011. Data were obtained from the VicRoads CrashStats, a publicly accessible crash statistics and mapping program. All crashes that involved collision between a cyclist and a vehicle door were extracted for analysis. Only crashes involving an injury to at least one person are reported by police in Victoria. Variables extracted for analysis included: gender, age, speed limit, road conditions, vehicle occupant (driver and passenger) and injury outcome. Salient details of this crash type were absent from some police reports. It was not possible to determine from all reports if the road user was entering or exiting the vehicle, or if the vehicle was stationary or parked. The available data was included in the analysis. 2.2. Hospital emergency department presentation data Cyclist crash presentations were analysed for the period from January 2000 to December 2010. Data were obtained from the Victorian Emergency Minimum Dataset (VEMD) in summary tables from the Victorian Injury Surveillance Unit (VISU) at the Monash Injury Research Institute (MIRI). The VEMD is a reliable and valid database that captures more than 80 per cent of all emergency department presentations at major trauma hospitals in Victoria (Stokes et al., 2000). All data related to patients whose injuries resulted from a cyclist-open vehicle door was extracted for analysis and were matched on the criteria ‘pedal cyclist rider/passenger’ and ‘door’ in the text narrative. Variables extracted for analysis included: age, gender and injury outcome. Details of the road users’ action prior to the event, including whether they were exiting or entering the vehicle or the vehicle was stationary or parked was not available. 2.3. Naturalistic study data A naturalistic cycling study of commuter cyclists was conducted from October 2009 to April 2010 in Melbourne, Victoria. In the 2.4. Data analysis Summary descriptive data are presented for the police and hospital data. The rate of cyclist-open vehicle door crashes was analysed. Calculating cyclist crash and/or injury rates in Australia is somewhat problematic due to the limited exposure data available for the number of people cycling. Nonetheless, cyclist participation figures used were from an annual national report generated by the Australian federal government on participation in exercise, recreation and sport activity (ERASS 2002–2011). This national telephone survey provides information on the number of people aged 15 years and over who have participated in a wide range of physical activities over the previous 12 month period. The ERASS reports are more representative than other Australian cycling indicators, such as the household census which is generated in winter and only refers to a single day of travel (Australian Bureau of Statistics, 2007) or population data which has no relationship to changes in the number of people cycling (Sikic et al., 2009). The video footage analysis was conducted in two stages, first the video was reviewed and the variables coded were: time of day (AM/PM); total trip time; number of parallel parked cars passed, and number of car door related events. Second was an in-depth analysis of the car door related events and variables coded were: event severity (collision, near-collision, incident). Definitions of the three categories of events are: collision is contact between a cyclist and another road user that involved a transfer of kinetic energy; near-collision is an event that requires rapid, evasive manoeuvring from the cyclist and/or the other road user to avoid a collision (e.g. hard braking or swerving); incident is an event in which some collision avoidance was required, but was less severe than the near-collision event. Type of event (driver entering vehicle, driver exiting vehicle, accessing rear driver-side door, and unknown, door partially/wide open); door opening (fully open/partially open); observed driver head check; driver reaction; vehicle type; pre-event cyclist behaviour; cyclist manoeuvre; cyclist reaction; cyclist post-event control; vehicle in adjacent lane; cyclist head check right, and visual obstruction. The variables were summarised with descriptive statistics for each event. From the video footage from the naturalistic cycling study, it was evident incidents occurred when the vehicle occupant was exiting or entering a vehicle. However, this data was not included in the police data or hospital data so it is not possible to analyse the actual crash data by the type of vehicle occupant behaviour (i.e. exiting or entering the vehicle). 137 M. Johnson et al. / Safety Science 59 (2013) 135–140 3. Results 3.1. Police reported crashes The majority of cyclists involved in crashes were aged 18 years or older (97.5%). Most crashes occurred at a midblock section of road (74.3%) and crashes were more likely to occur during the week (82.1%) than on the weekend (17.9%). Most crashes occurred in daylight hours (81.2%) and during peak travel times (8–10 am: 18.7%; 5–7 pm: 22.6%). Most crashes occurred in urban environments in Melbourne, excluding the central business district (71.4%) and in speed zones up to 60 kph (93.1%). Fig. 4 presents crashes by speed zone. The majority of collisions occurred in speed zones of up to 60 kph which is not surprising, given that many urban roads in Melbourne are zoned at these speeds. The number of collisions in lower speed zone areas, particularly 50 km/h, increased during the study time period and may be partly an artefact of the reduction of the default urban speed limit from 60 kph to 50 kph plus the introduction of 40 km/h in areas of high pedestrian and cyclist activity. From the police reported crashes, the majority involved the driver opening their door to exit or enter the vehicle (79.4%). However, 20.3 per cent involved a passenger opening their door. A significant proportion of the passengers opened the door of a vehicle that was not parked, but was stationary in the traffic lane. 3.2. Hospital emergency department presentation data All cyclists involved in a crash with a vehicle door who presented to hospital were aged 15 years or older; the greatest pro- Fig. 1. Total number of police reported crashes of cyclists-opened vehicle door, Victoria, 2000–2011. Count The total number of police reported cyclist crashes from January 2000 to December 2011 was 14,888, of these crashes, a total of 1247 (8.4%) involved a parked or stationary vehicle door. The total number of hospital presentations of people who had crashed while riding a bicycle from January 2000 to December 2010 was 12,844, of these crashes, a total of 401 (3.1%) involved a parked or stationary vehicle door. The significant difference between the police reported cyclistdoor crashes (n = 1247) and hospital presentations (n = 401) may be accounted for by the severity of injuries sustained by the cyclist. According to the police reported crashes, the majority of crashes were ‘other injuries’ (72%) with ‘serious injuries’ accounting for less than a third (27%) and one fatal crash. It is likely that the serious injuries are the most likely to present to hospital, which correlates more closely with the hospital data (police, serious injuries, n = 342). However, there was a dramatic increase in the number of cyclist–vehicle door crashes at occurred from 2008 to 2009 according to both police data (81.7% increase) (Fig. 1) and hospital data (77.8% increase) (Fig. 2). The increases were maintained in years that followed. This sudden increase in cyclist–vehicle door crashes was not reflected in the total number of cyclist crashes reported for that period (police increase: 2.2%; hospital increase 6.1%) (Fig. 3). 80 60 40 20 0 Year Fig. 2. Number of hospital presented cyclists following a collision with an opened vehicle door, Victoria, 2000–2010. Fig. 3. Total number of cyclist crashes, excluding cyclist–vehicle door (police and hospital data). portion was aged 20–29 years (42.1%), or 30–39 years (28%). The majority of cyclists were male (65.8%). Most injured were sustained to cyclists’ upper extremity (45.6%), lower extremity (16.7%), head/face/neck (16%) and multiple body regions (14.5%). The shoulder was the most commonly injured specific body region (16%). Main injuries sustained were superficial (23.7%), open wound (19.2%), fracture (18.5%), sprain or strain (16.5%). A small number of cyclists (n = 7, 1.7%) sustained an intracranial injury. The majority were treated in the emergency department and discharged home (83.8%), with 16.2 per cent admitted to hospital. 3.3. Rate of cyclist-opened vehicle door crashes It is important to contextualise the raw crash numbers in the context of cycling exposure. Cycling participation has increased in Victoria from 2000 (461,200) to 2010 (555,200) (increase 20.3%). However, the population has also increased and proportionally, the cycling participation increase is not significant (2000: 12.1%; 2010: 12.7%) (Department of Communications Information Technology and the Arts, 2002, 2011). The number of police reported and hospital presentations of cyclist-opened vehicle door crashes was standardised to determine the rate per 100,000 cycling participants (Fig. 5). 3.4. Naturalistic study data From the naturalistic cycling study data, 13 h and 33 min of video footage, that included 22 trips from 11 participants, was analysed. In total, 13 car door related events were identified (for examples of events, see video clips). The participants rode on various types of roads and parking facilities including narrow inner-city residential roads with vehicles densely parallel parked, clearways with no parked cars, and multi-lane arterial roads with parallel parking. A summary of the time analysed, the number of parallel parked vehicles passed and the events by gender, is presented in Table 1. In total, cyclists passed by 6197 parked vehicles, across an average trip (38 min), this equates to a cyclist passing a parked vehicle every 8 s. Of the vehicles passed, 13 doors were opened into the path of participants. No cyclists collided with a vehicle door, however 1 rider (female) had a near-collision that required her to brake heavily. The number of opened vehicle doors per trip ranged from 0 to 3 with a per trip rate of 0.59. The rate of cyclist-opened door 138 M. Johnson et al. / Safety Science 59 (2013) 135–140 Fig. 4. All police reported cyclist-opened door crashes by speed zone, Victoria, 2000–2011. Police crashes Rate of crashes per 100,000 cycling participants Hospital presentations Year Fig. 5. The rate of cyclist-open vehicle door crashes (per 100,000 cycling participants), Victoria, 2000–2010. Source: Participation data: ERASS, 2002–2011, Police reported crashes: CrashStats, 2012, Hospital presentations: VISU, 2011. events was related to the number of parallel parked vehicles overtaken. In all events, a collision was avoided by the cyclist’s evasive action. A list of descriptive variables of the events is included in Table 2. All vehicle occupants/drivers who were observed entering the vehicles opened the door. None of the vehicle occupants/drivers made a head check in the direction of the oncoming traffic and none of them waited for the cyclist to pass before opening the door. All vehicles observed were non-commercial vehicles. For 3 events, the cyclist approached a vehicle with the door partially or completely opened, however no person was filmed in the vicinity of the vehicle. 4. Discussion Cyclists are vulnerable road users and impacts with parked vehicles, especially when car doors are opened in front of travelling cyclists, are a significant issue for cyclist safety. While it is difficult to determine the true extent of this crash type, the findings revealed that cyclist-open vehicle door crashes are almost exclusively an adult cyclist crash type, and cyclists have relatively high exposure to this potential crash type. While riding alongside parallel parked vehicles, cyclists need to be highly attentive and able to react quickly to avoid a collision with an unexpectedly opened, often flung open, vehicle door. In this study, on an average trip, cyclists passed a parallel parked vehicle every eight seconds. Each time a cyclist passed a parked vehicle they needed to check to see if the vehicle is occupied, about to leave a parking bay or whether an occupant is about to leave or enter the vehicle. The vigilance required for cyclists to, almost constantly, assess the potential threat as they pass parallel parked vehicles requires a high level of concentration. Concentration that increases the task demand on cyclists and may reduce their capacity to assess other potential hazards. Avoiding an opening door requires complex and rapid reaction. Cyclists need to swerve to avoid the opening door while simultaneously making ahead check to assess surrounding traffic and depending on the cyclists’ travel speed, they may also need to brake, adding to the difficulty of bike handling. From the video footage analysed, drivers rarely head checked in the direction of the cyclist and it remained for the cyclist to react to avoid a collision. It is likely that cyclists’ evasive actions contribute to the number of cyclist-open door crashes and injuries not being much higher than recorded. A significant gender difference was identified in the police and hospital data and the video footage with more males experiencing crashes and having higher exposure to parked vehicles than females. Several factors may contribute to this, from the video footage it was evident that females rode passed fewer parked vehicles. This may be a conscious route selection choice that reduced females’ exposure. Cyclist speed may also be a factor, although speed was not recorded in this study. Differences in attentiveness and caution between males and females may also be factors. These findings have implications for the safety of cyclists and these are discussed within the Safe System Framework that underpins road safety in Australia including: safer roads and roadsides, safer road users, safer vehicles and safer speeds, and are discussed below. 4.1. Safer roads and roadsides It is likely that the most effective countermeasure would be a reconfiguration of the road space to shift cyclists away from the Table 1 Summary of time analysed, number of parked vehicles passed and events, by gender. Time of day All participants(n = 11) Gender Males (n = 8) Females (n = 3) AM PM AM PM Total trip time (hours:mins) Trip time in minutes min–max (average trip in minutes) No. parked parallel vehicles passed (average per trip) Events 13:46 18–59 (38) 6197 (281) 13 4:48 5:01 1:55 2:01 18–52 (36) 21–59 (37) 21–51(38) 25–55 (40) 2185 (273) 2501 (312) 485 (161) 511 (170) 2 10 1 0 M. Johnson et al. / Safety Science 59 (2013) 135–140 4.2. Safer road users Table 2 Summary of in-depth analysis of cyclist-car door events. Total number of events Event severity Incident Near-collision Type of event Driver entering vehicle Accessing rear driver-side door Unknown, door partially/wide open Driver exiting vehicle Door opening Fully open Partially open Observed driver head check No Unknown Driver reaction No reaction Unknown Vehicle type Sedan 4WD Hatchback Station wagon Pre-event cyclist behaviour Safe and legal Cyclist manoeuvre Riding straight Cyclist reaction Veered wide Stopped Cyclist post-event control Maintained control Vehicle in adjacent lane Yes Cyclist head check right Yes Visual obstruction No 139 n = 13 12 (92.3%) 1 (7.7%) 5 3 3 2 (38.5%) (23.1%) (23.1%) (15.3%) 9 (69.2%) 4 (30.8%) 11 (84.6%) 2 (15.4%) 11 (84.6%) 2 (15.4%) 7 3 2 1 (53.8%) (23.1%) (15.4%) (7.7%) 13 (100%) 13 (100%) 12 (92.3%) 1 (7.7%) 13 (100%) 3 (23.0%) 1 (7.7%) 13 (100%) door zone. The most common and widely implemented cyclist facility in Australia is the bike lane, particularly in metropolitan areas. The recommended bike lane widths are set by a national standard with the minimum acceptance of 1.2 m in 60 kph zones (Austroads, 2011). However, in practice, bike lane widths vary, depending on the available space after vehicle travel lanes and parking bays are accommodated and are often too narrow for the cyclists to safely swerve away from the opening door and stay within the boundaries of the bike lane. If the standard bike lane configuration is to be safe for cyclists, the minimum width needs to be sufficient for cyclists to safely avoid opening doors and stay within the boundaries of the bike lane. 4.1.1. Safer roads and roadsides – reconfiguration of the road design At some inner-city areas of Melbourne, the vehicle-bike lanes have been reconfigured with the bike lane moved from the outside of the parallel parking bay, to between the curb and the parking bays. This configuration moves cyclists away from the moving vehicular traffic and the risk of opening doors on the drivers’ side of the vehicle while the parked vehicles create a buffer between the vehicular lane and the bike lane. At some locations, concrete curbs or vertical posts create a safety buffer for cyclists. However, this configuration does increase cyclist-pedestrian interactions as now all vehicle occupants must cross the bike path to reach the footpath. A comprehensive review of current parking options and the potential alternative use of existing spaces are needed to maximise vehicular movements, parking, as well as pedestrian and cyclist safety. 4.2.1. Safer road users – vehicle occupants Driver behaviour was a major contributing factor in the cyclistopen vehicle door crash type. The majority of drivers apparently did not look before opening their vehicle door and none of the drivers/vehicle occupants stopped to wait for the cyclists to pass before opening the vehicle door. An obvious and direct solution is to ensure drivers always check for cyclists before opening any vehicle door. However, the current practice of not checking for cyclists is most likely habitual for many drivers and to change this behaviour is likely to require extensive investment through behaviour change programs and may or may not be adopted by all drivers all the time. One way to stop such habits from forming is through training and education programs for new drivers that include safe interactions with cyclists. For new drivers, additional information and testing in the drivers’ licence test could include safely entering and exiting the vehicle, including a head check. In addition, practices from countries with high cycling participation rate such as Denmark and the Netherlands could be adopted (e.g., van Leeuwen, 2006). Such examples may include when parallel parked: opening door with ‘other’ hand (left in Australia) to encourage drivers to twist in their seat and head check before opening the door; opening the door as little as necessary when entering and exiting the vehicle; all passengers exit on the curbside, and approach the vehicle by walking towards traffic. Cyclist collisions with opened vehicle doors have added complication of crash risk from the traffic in the adjacent lane (Savage, 2011; Spooner, 2011). Drivers need to allow adequate lateral distance when travelling parallel to cyclists. Currently in Australia, there is no legal requirement dictating the clearance distance a driver needs to provide when overtaking a cyclist. Recommendations range from at least 1 m (1 m) (VicRoads, 2010) to 2 m in higher speed zones (over 70 km/h) (Department of Transport, 2010). The Amy Gillett Foundation (2009) has introduced an education/ awareness campaign, ‘A Metre Matters’ to address this issue. The current law in Victoria is that drivers must not cause a hazard to another road user, including cyclists, when opening their door, with associated monetary penalties (VicRoads, 2012). However, there are examples that the current legislation is not being enforced following crashes (Spooner, 2011). If the existing laws regarding driver behaviour are not currently effective or enforced, it is essential that these laws are reviewed to ensure that cyclists are judicially protected. Improved vehicle passenger behaviour is also important to reduce this crash type. In particular, taxi passengers were a significant proportion of passenger door crashes. Greater education/ awareness campaigns are needed and enforcement options such as monetary penalties or demerit points for passengers need to be considered to improve behaviour. 4.2.2. Safer road users – cyclists For cyclists, two specific behaviours are important to avoid a collision with an opened vehicle door: travel at a speed within bike handling skills to ensure control of the bike including safe braking or direction change, and vigilance about drivers/passengers opening car doors, check for heads through the back window, side mirrors. In addition, cyclists’ use of front lights in the twilight periods at dawn and dusk and during days that are overcast/poor visibility may increase their conspicuity to drivers who do make a head check. 4.3. Safer vehicles To date, the majority of vehicle technology and safety features have focused on vehicle occupant functions, vehicle occupant 140 M. Johnson et al. / Safety Science 59 (2013) 135–140 protection and some pedestrian detection technologies. Comparatively little emphasis of vehicle technology and safety features has focused on the side or rear of the vehicle. Existing blind spot warning technologies could adapted to detect cyclists even when the vehicle is stationary and activated as the door is about to be opened. Delayed door opening or signal lights when a door is about to be opened would give cyclists an opportunity to see the opening door and increase their time to react safely. Cyclist safety needs be included in the range of tests conducted by the Australasian New Car Assessment Program (ANCAP) program to improve cyclist safety in all types of cyclist–vehicle collisions. Blind spot warnings could be given greater weight in the testing if they could be shown to prevent doorings. 4.4. Safer speeds The importance of speed in this crash type relates to the cyclist and the drivers in the adjacent lane. Cyclists need to travel at speeds that are within their bike handling skills and understand the braking distance required. In relation to drivers’ speed, the majority of the reported crashes occurred in speed zones up to 60 kph (94.1%). One solution may be to reduce vehicle speed to minimise the injury severity if a crash occurs, while simultaneously giving drivers more time to react. 5. Strengths and limitations The strength of this study was in the use of three complementary data sources. While there were discrepancies in the number of reported crashes from police and hospital data, the data provided insights into the details of crashes and crash outcomes. The use of the naturalistic video footage provided unique data of the exposure of cyclists to the risk of opened vehicle doors. The naturalistic data provided a clear view of the vehicle as the cyclist approached. The 180° scope of the camera captured all vehicle occupants as the cyclist approached. However, there was limitations in the visibility of vehicle occupants as they exited the vehicle. It was not possible from the perspective of the helmet camera, and therefore potentially the viewpoint of the cyclist, to determine if the vehicle occupant had made a head check prior to opening the vehicle door. The limitations of the study relate to the poor data available, it would appear that this crash type is somewhat under-reported and it was difficult from the hospital and police data to gain a comprehensive picture of the extent of this crash type. In addition, it is likely that the speed of the cyclists may contribute to their exposure to this crash type; the lack of speed recordings in the video footage was a limitation. Future studies that recorded speed data would enhance the understanding of the role of cyclist speed in this crash type. 6. Conclusion Cyclist-open vehicle door crashes result in a significant and growing number of cyclist injuries each year. Naturalistic cycling observational data shows the extent of these collision types is likely to be mediated by the attention and bike handling skills of the cyclists, rather than the drivers’ behaviour or the current cycling environment on the road. The most significant improvement may come from driver focused behaviour awareness and change programs, however there is also an important role for the continued implementation of effective cycling facilities that accommodate cyclists, either by physically separating them from traffic, or ensuring an adequate space to safely avoid drivers who open vehicle doors in their path. Acknowledgements The authors thank Angela Clapperton and Emily Kerr from VISU for their assistance with the hospital data extraction and review. The naturalistic cycling study was undertaken as part of a doctoral research project (M. Johnson). The authors thank the sponsors for their support of the doctoral research through a joint scholarship from the Monash University Accident Research Foundation and the Amy Gillett Foundation Safe Family Research Scholarship provided by Bradley Bayly Legal. This publication is supported by the Monash University Postgraduate Publications Award (M. Johnson). The authors thank the University for providing support through this award. References Amy Gillett Foundation, 2009. A metre matters. <http://www.amygillett.org.au/ametre-matters>. Atkinson, J.E., Hurst, P.M., 1983. Collisions between cyclists and motorists in New Zealand. Accident Analysis and Prevention 15 (2), 137–151. Australian Bureau of Statistics, 2007. 20680 – Method of Travel to Work (full classification list) by Sex – Australia (Retrieved 20.12.08). Austroads, 2011. Cycling Aspects of the Austroads Guides. Sydney. Publication No. AP-G88/11. City of Melbourne, 2004. Issue paper: Cycling. <www.melbourne.vic.gov.au/rsrc/ PDFs/MelbourneTransportStrategy/MelbTS_Cycling1407.rtf>. Daff, M., Barton, T., 2005. Marking Melbourne’s Arterial Roads to Assist Cyclists. ITE 2005 Annual Meeting and Exhibit. Institute of Transportation Engineers. Melbourne, Australia, <http://209.85.173.104/search?q=cache:OxpPBB8gK4AJ: www.ite.org/Conference/SamplePaper.pdf+%22Advanced+stop+lines+for+ cyclists%22&hl=en&ct=clnk&cd=7&gl=us&client=firefox-a> (Viewed 10.07.08). Dennerlein, J.T., Meeker, J.D., 2002. Occupational injuries amongst Boston bicycle messengers. American Journal of Industrial Medicine 42, 519–525. Department of Communications Information Technology and the Arts, 2002. Participation in Exercise, Recreation and Sport, Annual Report 2001. Department of Communications Information Technology and the Arts, 2011. Participation in Exercise, Recreation and Sport, Annual Report 2010. Department of Transport, 2010. A Handbook for Western Australian Road Users. G. o. W. Australia. Fincham, B., 2006. Bicycle messengers and the road to freedom. The Editorial Board of the Sociological Review, 209–222. Haileyesus, T., Annest, J.L., et al., 2007. Cyclists injured while sharing the road with motor vehicles. Injury Prevention 13 (3), 202–206. Harman, B., 2007. Bicycle Safety and Law Enforcement. Australasian Road Safety Research Policing and Education Conference, Melbourne. Harris, S., 1990. The real number of road traffic accident causalities in the Netherlands: a year-long survey. Accident Analysis and Prevention 22 (4), 371–378. Johnson, M., Charlton, J., et al., 2010. Naturalistic cycling study: identifying risk factors for on-road commuter cyclists. Annals Advanced Automotive Medicine 54, 275–283. Pucher, J., Dijkstra, L., 2003. Promoting Safe Walking and Cycling to Improve Public Health: Lessons from the Netherlands and Germany. American Journal of Public Health 93 (9), 1509–1516. Savage, J., 2011. ‘Car door’ driver charged over cyclist’s death. nzherald.com.nz, Auckland. Sikic, M., Mikocka-Walus, A., et al., 2009. Bicycle injuries and mortality in Victoria, 2001–2006. Medical Journal of Australia 190, 353–356. Spooner, H., 2011. Inquest into the Death of James Bernard Cross. Melbourne, Coroners Court of Victoria. Stokes, M., Ozanne-Smith, J., et al., 2000. Validation of an injury surveillance epidemiological data system used within emergency departments. Injury Control and Safety Promotion 7, 219–232. Tomlinson, D., 2000. Conflicts Between Cyclists and Motorists in Toronto, Canada. Velo Mondial, World Cycling Conference. van Leeuwen, G., 2006. Theorieboek rijbewijs B. Vol. Translation by Miranda Cornelissen, 2010. ANWB Multi Mediale Uitgeverij, The Hague. VicRoads, 2010. Road to Solo Driving: Learning to Drive, Driving to Learn. Roads Corporation Victoria, Kew, Victoria. VicRoads, 2012. ‘‘CrashStats.’’ <http://www.vicroads.vic.gov.au/Home/ SafetyAndRules/AboutRoadSafety/StatisticsAndResearch/CrashStats.htm> (Retrieved 22.03.12). Wood, J.M., Lacherez, P.F., et al., 2009. Drivers’ and cyclists’ experiences of sharing the road: Incidents, attitudes and perceptions of visibility. Accident Analysis and Prevention 41 (4), 772–776.