CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES

CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES

CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES Narelle Haworth Peter Vulcan Lyn Bowland Nicola Pronk Monash University Accident Research Centre September 1997 Report No. 120

ii MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE REPORT DOCUMENTATION PAGE Report No. Date ISBN Pages 120 September 1997 0 7326 0700 0 xii + 52 + app Title and sub-title: Characteristics of fatal single vehicle crashes Author(s) N. Haworth, P. Vulcan, L. Bowland and N. Pronk Type of Report & Period Covered: General, 1995-1997 Sponsoring Organisation(s): This project was funded through the Centre s Baseline Research Program for which grants have been received from: Department of Justice Royal Automobile Club of Victoria (RACV) Ltd Roads Corporation (VicRoads) Transport Accident Commission Abstract: This report describes the characteristics of a set of 127 fatal single-vehicle crashes which occurred within 200 km of Melbourne during the period from 1 December 1995 to 30 November 1996. The crashes comprised the cases for the Case-control study of fatal single-vehicle crashes. The main identifiable factor contributing to the occurrence of fatal single vehicle crashes in this study was BAC>.05 (and particularly BAC>.15) and the main identifiable factors contributing to the degree of severity of the crashes appeared to be impacts with trees and poles and some older vehicles which are unlikely to comply with most of the current safety standards. A number of possible improvements to procedures for the investigation of serious injury crashes are recommended. Key Words: (IRRD except when marked*) Single vehicle accident, fatality, drink driving, marijuana, driver characteristics, vehicle, road environment, pole, vegetation. Reproduction of this page is authorised Monash University Accident Research Centre, Wellington Road, Clayton, Victoria, 3168, Australia. Telephone: +61 3 9905 4371, Fax: +61 3 9905 4363 CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES iii

iv MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

EXECUTIVE SUMMARY This report describes the characteristics of a set of 127 fatal single-vehicle crashes which occurred within 200 km of Melbourne during the period from 1 December 1995 to 30 November 1996. The crashes comprised the cases for the Case-control study of fatal single-vehicle crashes. The cases have location, driver/rider and vehicle characteristics. The controls are (non-crash) trips which also have location, driver/rider and vehicle characteristics. The comparisons of cases and controls to derive relative risk estimates are presented in the companion report entitled Estimation of risk factors for fatal single-vehicle crashes. A summary of the overall study has also been produced (Single vehicle crash study: Summary report). The aims of the Case-control study of fatal single-vehicle crashes study were to: 1. investigate single vehicle crashes to determine the circumstances and factors contributing to them 2. estimate the over-involvement (relative risk) of these factors 3. identify improvements in procedures for the investigation of road deaths and life threatening injuries 4. provide information from which countermeasures can be developed to the agencies responsible for road safety in Victoria This report addresses aims 1 and 3. CHARACTERISTICS OF THE CRASHES There were 127 single vehicle crashes in which 133 persons were killed. The crashes had the following characteristics: Type of crash almost 75% of crashes involved an impact with a tree or pole or both (71% of metropolitan, 78% of rest of study area crashes) tree crashes were more common outside the metropolitan area 53% of trees and 36% of poles impacted were in the median or on the righthand side of the road 31% of crashes involved impact with a tree or pole located within the desirable clear zone outlined in VicRoads policy Temporal characteristics crashes were most common from midnight to 6 am and midday to 6 pm crashes were most common on Saturdays and Sundays there were no clear seasonal effects CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES v

Persons killed five crashes were multiple fatalities the driver was killed in 81% of crashes almost half the persons killed were aged under 25 Drivers and riders 80% were male more than 40% were aged under 25 natural disease (particularly heart disease) was the cause of death for at least 60% of the drivers aged 60 and over 13% had been severely depressed in the previous six months 13% did not hold a current licence (6% never held a licence, 2% licence lapsed, 6% licence disqualified) Alcohol and other drugs BAC was known for 95% of crashes and cannabis known for 83% of crashes. The percentages in this section are percentages of known cases. All of the cases of cannabis being present were in deceased drivers where it was detected as the metabolite, carboxy-thc. This substance may be detected some weeks after cannabis use and therefore its detection does not prove that the driver was under the influence of cannabis at the time of the crash. BAC>.05 in 36%, BAC>.15 in 27% of crashes cannabis was most commonly found in conjunction with alcohol: for BAC>.05, cannabis was present in 49% for BAC>.15, cannabis was present in 52% but for BAC<.05, cannabis was present in only 6% males were twice as likely as females to have BAC>.05 but the proportions of these with BAC>.15 were similar for males and females no alcohol or cannabis was present in drivers aged 60 and over drivers aged 25 to 59 were more likely to have BAC>.15 than those under 25 higher levels of education were related to fewer drivers with BAC>.05 or cannabis for 25 to 59 year olds but not for under 25s proportions of crashes with BAC>.05 or cannabis were similar in metropolitan and rural areas BAC>.05 was found more often on major roads than freeways, highways or minor roads cannabis was detected for 23% of male drivers and 11% of female drivers cannabis was detected in 21% of those under 25 and 25% of those aged 25 to 59 cannabis-involved crashes were less common on highways than freeways, major roads or minor roads information about consumption of prescription and nonprescription drugs was available for 78% and 73% of drivers, respectively 17% of drivers had taken prescription drugs in the 12 hours prior to the crash (14% of under 25s, 18% of 25 to 59s and 27% of those aged 60 and over) 3 drivers or riders had taken nonprescription drugs in the 12 hours prior to the crash vi MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

Passengers 74 passengers of whom 30 were killed were more likely to be of the same sex as the driver than to be of the opposite sex or at least one of each sex young drivers were more likely to have passengers than other drivers most passengers were in the same age group as the driver of the vehicle (particularly for under 25s) in four of the 24 crashes for which BAC of both driver and (killed) passenger were known, the BAC of the driver was greater than that of the passenger Vehicles 10 crashes involved motorcycles, 7 involved light commercial vehicles and 3 involved trucks six of the 10 motorcycles had an engine capacity of greater than 500 cc but none of the motorcycles with an engine capacity greater than 260 cc was ridden by a learner or probationary rider almost 20% of the vehicles involved in crashes were manufactured before 1978 and so were not required to comply with some of the current safety standards age of the vehicle was not related to age of the driver Locations of crashes 60% of crashes occurred in Melbourne metropolitan area, 40% in the remainder of the study region most crashes occurred on highways (42%) a larger proportion of crashes occurred on highways and a smaller proportion on minor roads in the metropolitan area compared to the rest of the study area two-thirds of the crashes occurred on undivided roads (58% of metropolitan, 77% of rest of study area) half of the crashes occurred where the speed limit was 100 km/h or 110 km/h (23% of metropolitan, 82% of rest of study area) about one-third of the crashes occurred on curves (27% of metropolitan, 36% of rest of study area) 20% of metropolitan and 8% of rest of study area crashes occurred at intersections (most commonly T intersections) traffic controls were present at the site of 13% of crashes (18% of metropolitan, 6% of rural) mud, oil or loose material was present on the road surface for 10% of crashes in the rest of the study area the road had a shoulder for 62% of metropolitan crashes and 96% of rural crashes. In the metropolitan area, 49% of shoulders were sealed compared with 20% of rural shoulders POSSIBLE IMPROVEMENTS IN INVESTIGATION PROCEDURES In order to improve the availability of information about severe crashes, it was recommended that: CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES vii

1. The availability of toxicological information for drivers who may be charged be investigated and, if necessary, a procedure be developed to incorporate the BAC data into the State Traffic Accident Record at an appropriate time (perhaps after completion of criminal procedures). 2. Given the lack of information on drugs in nonfatal crashes, perhaps testing of all (or a given proportion) of blood samples for drugs should be undertaken for a specified period to increase our knowledge in this area. 3. Blood samples of all drivers in fatal crashes (whether or not injured) be taken and analysed for both alcohol and other drugs. An alternative would be to have blood samples taken and analysed whenever the breath shows alcohol (building on the strong relationship between alcohol and cannabis). 4. The family s right to object to autopsy be maintained but that the authorities have the right to take and analyse a blood sample from the deceased. 5. Testing for the active ingredient, rather than the metabolite, of cannabis be undertaken, at least for fatal crashes in the short term. Improvements in technology may reduce costs etc and allow this to be extended to nonfatal crashes in the future. 6. Vehicle Identification Numbers (VINs) be recorded to allow better identification of makes and models of vehicles and judgements of the presence of safety features. Currently, VINs could be added to the State Traffic Accident Record routinely by interrogating the registration data, using the vehicle registration number. In the future, automatic capture of VINs at the crash site may replace this method. 7. In addition, for the next couple of years, it would be desirable if deployment of airbags could be recorded. 8. It is probably preferable for items regarding the state of the vehicle immediately prior to the crash (e.g. heating and ventilation) to be collected in special studies, rather than to burden the Police with additional workload for every reported crash. 9. While resource limitations are an important issue which needs to be considered, it may be useful from the point of view of training Accident Investigation Section staff and developing a high quality knowledge base for prevention, to have a special focus on particular crash types (or perhaps even locations) for a particular period. viii MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

TABLE OF CONTENTS Page Acknowledgments xi 1. Introduction 1 1.1 Background to the study 1 1.2 Outline of the study 1 1.3 Aims of the study 2 1.4 Structure of reporting 2 2. Study design and procedure 3 2.1 Crash types included 3 2.2 Study region 3 2.3 Questionnaire 3 2.4 Roles of the agencies involved 4 2.4.1 Victoria Police Accident Investigation Section 4 2.4.2 State Coroners Office 5 2.4.3 Monash University Accident Research Centre 5 2.5 Case review meetings 5 3. General characteristics of cases 7 3.1 Type of crash 7 3.2 Temporal characteristics of the crashes 8 4. Persons involved 11 4.1 Persons killed 11 4.2 Demographics of drivers and riders 11 4.3 Licensing 14 4.4 Alcohol and other drugs 15 4.4.1 Toxicology reports 15 4.4.2 Alcohol and cannabis 16 4.4.3 Correlates of alcohol use 17 4.4.3.1 Sex of driver or rider 17 4.4.3.2 Age 18 4.4.3.3 Level of education 18 4.4.3.4 Location of crash 19 4.4.3.5 Licensing 19 4.4.4 Correlates of cannabis use 19 4.4.4.1 Sex of driver or rider 19 4.4.4.2 Age 19 4.4.4.3 Education 19 4.4.4.4 Location of crash 20 4.4.4.5 Licensing 20 4.4.5 Prescription and nonprescription drugs 20 4.5 Passengers 21 CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES ix

Page 5. Vehicles involved 23 5.1 Types of vehicles involved 23 5.2 Heating and ventilation 24 5.3 Sound 25 5.4 Vehicle defects 25 6. Locations of crashes 27 6.1 Type of road 27 6.2 Speed limit 28 6.3 Road geometry 28 6.3.1 Type of intersection 29 6.4 Delineation 29 6.5 Road furniture 30 6.6 Traffic controls and Local Area Traffic Management devices 31 6.7 The road surface 31 6.7.1 Road shoulder 32 6.8 Roadside objects 33 6.9 Environmental factors 33 7. Crashes involving trees and poles 35 7.1 Distances to impacted trees and poles 35 7.2 Comparisons with the VicRoads Clear Zone Policy 38 7.3 Crashes with trees and poles on curves 38 8. Summary of characteristics of crashes 41 9. Discussion 45 9.1 Alcohol and cannabis 45 9.2 Comparisons with other data sources 46 10. Possible improvements in investigation procedures 49 10.1 Alcohol 49 10.2 Other drugs 50 10.3 Recording vehicle data 51 10.4 AIS attendance criteria 51 References 52 Appendix 1: Crash questionnaire Appendix 2: Coding modifications to vehicles Appendix 3: Summary of cases Appendix 4: Concentrations of cannabis metabolite x MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

ACKNOWLEDGMENTS This study could not have been undertaken without the considerable efforts of the Victoria Police Accident Investigation Section and Traffic Operations Group and the State Coronial Services Centre. We acknowledge their dedication and commitment to the study. The technical input and support of the Project Steering Committee is appreciated. The members of the Committee were: Graeme Johnstone, State Coroner Isadora Engert - State Coroner s Office Bob Wylie, Neil O Sullivan, Noel Osbourne - Victoria Police Ted Vincent, Mike Regan, Mike Tziotis, Bob Gardner, Phillip Jordan - VicRoads David Healy - Transport Accident Commission Karen McIntyre, Peter Doupé, Anne Harris - RACV David Ranson - Victorian Institute for Forensic Medicine Many of our colleagues at MUARC provided valuable advice during the study. We would like to acknowledge the traffic engineering input of Bruce Corben and the statistical advice of Max Cameron, Lesley Day and Caroline Finch. Thank you also to the drivers and riders who agreed to be interviewed for the study. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES xi

xii MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

1. INTRODUCTION 1.1 BACKGROUND TO THE STUDY Single vehicle crashes resulted in 134 deaths and 1898 persons seriously injured in Victoria during 1995 (VicRoads, 1996). The monetary cost to the community of these types of crashes is estimated to be about $450 million. Overall, single-vehicle collisions comprise approximately 30% of road trauma in Victoria. The study arose from concern by the Victoria Police and the State Coroner that there was a need for a better understanding of the causative factors in single vehicle runoff-road crashes. Traditionally, Police investigation of these crashes has often been rudimentary when the driver has been killed (and so prosecution is not possible). Since driver-killed crashes comprise about three-quarters of all fatal single-vehicle crashes, the lack of detailed investigation of these crashes has prevented the development of a clear understanding of the factors involved. The Police and Coroner, in consultation with the Transport Accident Commission and Vic Roads, decided that there was a need for a multi-disciplinary approach which would go beyond the attribution of blame. The study was very much a cooperative venture between the State Coroner s Office, Victoria Police Accident Investigation Section, Victoria Police Traffic Operations Group and Monash University Accident Research Centre. In most past studies, crash types have been studied in a descriptive manner, focussing on in-depth investigation to identify those factors which contribute to crash occurrence or injury severity. The lack of a control group in these studies, however, has made it difficult to test whether the factors identified truly increase crash risk or are merely common characteristics of the road users, vehicles and locations involved. Increasingly, the epidemiological tool of the case-control study is being applied to investigation of road safety problems, as it is in medical research (e.g. cancer). Thus this case-control study aimed to identify the factors which contribute to an increased risk of occurrence of fatal single vehicle crashes, an important step in developing countermeasures to address the problem. 1.2 OUTLINE OF THE STUDY The Case-control study of fatal single-vehicle crashes collected detailed information about cases and controls. The cases are fatal single-vehicle crashes (or crash trips) which occurred during the period from 1 December 1995 to 30 November 1996. The cases have location, driver/rider and vehicle characteristics. The controls are (non-crash) trips which also have location, driver/rider and vehicle characteristics. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 1

The study collected three types of information: detailed descriptive information about the crash and the resultant injuries comparisons of the features of cases and controls, and general exposure information (gathered as part of the control collection process). 1.3 AIMS OF THE STUDY The aims of the study were: 1. to investigate fatal single vehicle crashes to determine the circumstances and factors contributing to them 2. to estimate the over-involvement (relative risk) of these factors 3. to identify improvements in procedures for the investigation of road deaths and life threatening injuries 4. provide information from which countermeasures can be developed to the agencies responsible for road safety in Victoria This report addresses aims 1 and 3. 1.4 STRUCTURE OF REPORTING This report describes the characteristics of the crashes which comprised the cases for the Case-control study of fatal single-vehicle crashes. The comparisons of cases and controls to derive relative risk estimates are presented in the companion report entitled Estimation of risk factors for fatal single-vehicle crashes. A summary of the overall study has also been produced (Single vehicle crash study: Summary report). 2 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

2. STUDY DESIGN AND PROCEDURE 2.1 CRASH TYPES INCLUDED This phase of the study included only fatal single-vehicle crashes (rather than fatal and serious injury crashes) to limit the demands on Police resources and to enable Police involvement in the study to be maintained for a full year. Fatal single-vehicle crashes were defined in the following terms for the purposes of this study - where a collision occurs involving a single vehicle, resulting in the death of occupants of that vehicle, in either (sic) of the following circumstances: 1. collision with a fixed object (on/off road) 2. collision with parked vehicle 3. rollover collisions. This definition differs from that used by Vic Roads in the State Traffic Accident Record in several ways: it does not include crashes where a vehicle impacts a pedestrian it does not include crashes involving falling from or in a moving vehicle it includes crashes in which a moving vehicle impacts a stationary vehicle. The study included all fatal single-vehicle crashes in the study region, regardless of the type of vehicle involved. 2.2 STUDY REGION The study region comprised the area in Victoria within a 200 km radius of Melbourne. It stretched from Sale in the east, to Wangaratta and Echuca in the north, to Ararat in the west (see Figure 2.1) and included most of the major population centres in Victoria. 2.3 QUESTIONNAIRE The crash questionnaire was developed based on questionnaires used in the University of Auckland Case-Control Study of Motorcycle Crashes (personal communication) and input from the Project Steering Committee. A copy of the questionnaire used in this study is included as Appendix 1. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 3

Figure 2.1. The study region. MUARC provided blank crash questionnaires to the Victoria Police Accident Investigation Section (AIS), individually ring bound to keep all of the contents together and to make them more easily identifiable. The sections were colour-coded to facilitate completion. A manual for the crash questionnaire was developed by MUARC in consultation with the AIS and issued to each AIS team. The questionnaire for each case was completed by the investigating officer. Following discussion with Vic Roads Vehicle Safety Standards section, a set of categories were developed for the question, Have any modifications been made to the vehicle?. The coding guidelines are presented in Appendix 2. 2.4 ROLES OF THE AGENCIES INVOLVED The study was very much a cooperative venture between the State Coroner s Office, Victoria Police Accident Investigation Section, Victoria Police Traffic Operations Group and the Monash University Accident Research Centre. The roles played by these agencies in the research are outlined below. 2.4.1 Victoria Police Accident Investigation Section Members of the AIS attended the scene of the crash and conducted the investigations necessary to complete the crash questionnaire. This included conducting inspections of the site of the crash. The AIS Coordinator collected questionnaires from attending officers and forwarded them to MUARC. 4 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

The AIS Coordinator faxed notifications of crashes to the Coronial Services and MUARC each weekday. A special series of job numbers, prefixed with SV, was used for the study. The notification sheets included the following items: job number, date of crash, date of death (if delayed), time of crash, location of crash, nature of crash (e.g. struck pole, struck tree, rollover), AIS member attending, vehicle make and model, year of manufacture, name and address of deceased. In addition to collecting information about crashes, the AIS Coordinator notified the Traffic Operations Group of times and places where their assistance would be needed in the collection of controls (random breath testing and licence checking). The draft crash questionnaire was piloted by AIS members during late November 1995. A number of important modifications were made on the basis of AIS input. These included changes to wording and addition of some questions. 2.4.2 State Coroners Office The Coroners Office coordinated information for fatalities. This included the supply of toxicology reports and autopsy reports when available. While some post mortems were conducted in the country, all country autopsies were overviewed by a pathologist at the Victorian Institute of Forensic Medicine. Inquests were held in the city or the country. Even if there was a Chambers finding rather than an inquest, the data was available to the study. Further inquests may be held up to six months after the end of the study period. Criminal cases may be delayed for up to 18 months or two years. 2.4.3 Monash University Accident Research Centre MUARC designed and produced the questionnaires and undertook data entry and analysis. 2.5 CASE REVIEW MEETINGS In the initial project discussions, it was decided that at least some of the cases would be reviewed by a panel of experts to better identify the factors that contributed to the occurrence or severity of the crashes and also the manner in which they related to each other. It was also expected that the identification of these factors would provide the opportunity to develop better targeted countermeasures for single vehicle crashes. Case review meetings were held in March 1996, June 1996 and February 1997. Seven cases were reviewed during the first meeting. The countermeasure issues which were raised related to the efficacy of the medical review system for driver licensing, the possibility of educating older groups and doctors about fitness to drive, guardrailing of bridge pillars, terminal design for guardrails and educating particular ethnic groups about their susceptibility to alcohol. Other issues discussed included the role of natural disease in crashes and the statistical classification of these crashes. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 5

Six cases were presented by the AIS at the second meeting and seven cases at the third meeting. The issues discussed included fatigue caused by driving too long or prescription drugs, safety of driving by overseas visitors, the differences in the effects of alcohol and cannabis, detection of cannabis, airbag effectiveness, appropriateness of speed zoning at some locations and the rural culture of drink driving. While those attending found the case review meetings to be useful, the substantial resources needed to compile the information and hold meetings was acknowledged. It was considered not feasible to review every crash in the study. The Coroner suggested that, because of the time involved, only cases of special interest should be reviewed. 6 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

3. GENERAL CHARACTERISTICS OF CRASHES This chapter describes the characteristics of the 127 fatal crashes included in the study. Six additional cases were notified to MUARC but were later withdrawn from the study by the Victoria Police because they did not fit all of the selection criteria for inclusion in the study (because they were out of area, multi-vehicle or pedestrian etc). Appendix 3 summarises the general characteristics of the cases. 3.1 TYPE OF CRASH The following classification of crash types utilises the description provided on the Police notification form. The most common type of crash was an impact with a tree, both in Metropolitan Melbourne and the rest of the study area. As Table 3.1 shows, crashes with trees comprised almost 40% of crashes in Metropolitan Melbourne and more than 60% of crashes in the rest of the study area. Crashes with poles were the second most common type of crash overall, but were much more common in Metropolitan Melbourne (29%) than in the rest of the study area (10%). Rollovers were more common outside of the metropolitan area than in Melbourne (10% versus 4%). A number of the crashes were simply described as off-road on the notification form. Table 3.1. Percentages of crashes according to the type of crash. Crash type Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) Impact with pole 29 10 21 Impact with tree 39 63 49 Impact with stationary 5 0 3 vehicle Impact with bridge pylon 3 0 2 Impact with fence 4 2 3 Impact with railing 3 0 2 Impact with embankment 0 4 2 Rollover 4 10 6 Off road 4 10 6 Other 9 2 6 Total 100 101 100 There was evidence of braking for 20 crashes (16%). This comprised rear tyre hooked skid (5), rear tyre weaved skid (fishtailing) (5), straight skid (7) and front tyre skid (3). In two cases the evidence of braking was in relationship to a sudden change in the road surface. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 7

3.2 TEMPORAL CHARACTERISTICS OF THE CRASHES The months and the days of the week when the crashes occurred are summarised in Tables 3.2 and 3.3. Crashes were more common in February (14%), April (12%) and September (13%) than in the other months. There does not appear to be a clear seasonal trend in the crashes. Saturday (23%) and Sunday (20%) were the most common days for crashes to occur, with crashes being least frequent on Tuesdays (see Table 3.3). Table 3.2. The months in which the crashes occurred. Month Number of crashes Percent of crashes December 10 8 January 12 9 February 18 14 March 6 5 April 15 12 May 8 6 June 9 7 July 8 6 August 11 9 September 16 13 October 5 4 November 9 7 Total 127 100 Table 3.3. The days of the week that the crashes occurred. Day of the week Number of crashes Percent of crashes Monday 17 13 Tuesday 9 7 Wednesday 14 11 Thursday 15 12 Friday 18 14 Saturday 29 23 Sunday 25 20 Total 127 100 More crashes occurred at night (6 pm to 6 am) than during the day (6 am to 6 pm, 54% vs 47%). Crashes were most common from midnight to 6 am and from midday to 6 pm and least frequent from 6 am to midday (see Table 3.4). Figure 3.1 presents the time of occurrence in greater detail. A total of 45 crashes (35%) occurred in the six hour period between 10 pm and 4 am. The crash frequencies were generally highest from 2 am to 4 am and 10 pm to midnight. Many of the crashes between 2 am and 4 am occurred on Friday, Saturday or Sunday mornings (see Figure 3.2). 8 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

Table 3.4. The times of day that the crashes occurred. Time of day Number of crashes Percent of crashes Midnight to 6 am 39 31 6 am to midday 23 18 Midday to 6 pm 36 28 6 pm to midnight 29 23 Total 127 100 Figure 3.1. Number of crashes during each two-hour period of the day. 18 16 14 Number of crashes 12 10 8 6 4 2 0 0-2 2-4 4-6 6-8 8-10 10-12 12-14 14-16 16-18 18-20 20-22 22-24 Time of day Figure 3.2. Percentages of crashes during each two-hour period of the day separately for weekdays and weekends. Weekends are Friday, Saturday, Sunday. 25 Percent of crashes 20 15 10 5 Weekday Weekend 0 0-2 2-4 4-6 6-8 8-10 10-12 12-14 14-16 16-18 18-20 20-22 22-24 Time of day CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 9

10 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

4. PERSONS INVOLVED 4.1 PERSONS KILLED A total of 133 persons were killed in the 127 crashes. The five multiple fatality crashes comprised three in which the driver (or rider) and one passenger were killed, one in which the driver and two passengers were killed and one in which two passengers were killed (see Table 4.1). The driver (or rider) was killed in 103 crashes (81%). Table 4.1. Persons killed in fatal single vehicle crashes. Persons killed Number of crashes Percent of crashes Driver (or rider) only 99 78 One passenger only 23 18 Driver (or rider) and one passenger 3 2 Driver and two passengers 1 1 Two passengers 1 1 Total 127 100 Overall, 74% of those killed were male (81% of drivers or riders, 53% of passengers). Table 4.2 presents information about the ages of the persons killed. About 45% of the persons killed in the crashes were aged under 25. About 12% of the persons killed were aged 60 or over. Table 4.2. Age groups of people killed in fatal single vehicle crashes. Age group Drivers and riders Passengers All occupants Under 25 39 21 60 25 to 59 51 6 57 60 and over 13 3 16 Total 103 30 133 4.2 DEMOGRAPHICS OF DRIVERS AND RIDERS Overall, 80% of the drivers and riders in the crashes were male. The ages of the drivers and riders involved in the crashes are summarised in Figure 4.1. More than 40% of the drivers were aged under 25. Less than 12% of the drivers were aged over 60. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 11

Figure 4.1. Age groups of drivers and riders involved in fatal single vehicle crashes. 45 40 Percent of crashes 35 30 25 20 15 10 5 0 Under 25 25 to 34 35 to 44 45 to 54 55 to 64 65 to 74 75 and over Age group Overall, 54% of the drivers or riders were single (see Table 4.3). Not surprisingly, a larger proportion of the youngest group were single and only the oldest group contained widows (including widowers). The marital status of four drivers or riders had changed during the last 12 months: one under 25 year old had married, two 25 to 59 year olds had separated and one 25 to 59 year old had divorced. Table 4.3. Percentages of drivers or riders according to marital status. Marital status Age group Under 25 (n=52) 25 to 59 (n=60) 60 and over (n=15) Total (n=127) Single 94 30 7 54 Married or living with 6 53 53 34 partner Separated or divorced 0 13 0 6 Widowed 0 0 33 4 Unknown 0 0 7 2 Total 100 100 100 100 The highest level of education most commonly received by drivers or riders was secondary school (see Table 4.4). Not surprisingly, University education was most common among the youngest group. None of the drivers aged 60 and over were in full-time work, the majority were receiving a benefit (see Table 4.5). About half of 12 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

the drivers or riders in each of the other two age groups were in full-time work. Almost one-quarter of the under 25s were students. Table 4.4. Percentages of drivers or riders according to highest level of education received. Level of education Age group Under 25 (n=52) 25 to 59 (n=60) 60 and over (n=15) Total (n=127) Primary school 0 0 13 2 Secondary school 65 68 47 65 TAFE 6 3 7 5 University 21 7 7 13 Unknown 8 22 27 17 Total 100 100 101 102 Table 4.5. Percentages of drivers or riders according to present job situation. Individuals can belong to more than one category so totals may exceed 100%. Job situation Age group Under 25 (n=52) 25 to 59 (n=60) 60 and over (n=15) Total (n=127) Full-time work 50 53 0 46 Part-time work 12 10 0 9 Receiving benefit 13 13 67 20 Student 23 0 0 9 Self employed 0 12 0 6 Home maker 0 5 7 3 Unemployed and no 6 7 0 6 benefit Other 2 0 20 3 Unknown 2 5 7 4 Whether the driver or rider worked shifts was known for 83/127 crashes. Fifteen drivers or riders worked shifts, six who were aged under 25 and nine who were aged 25 to 59. The number of days since the driver or rider had a day off was known for 55 crashes. Only two of these drivers had worked for more than seven days since their last day off. The variable was not coded for those who were unemployed or receiving a pension or benefit. Whether the driver or rider had felt severely depressed at any time in the last six months (e.g. required treatment or attempted to take his/her life) was recorded for 109 of the 127 crashes. Fourteen of the drivers (13% of known) had been severely depressed. Fifteen drivers were reported to have had a disability. The disabilities recorded included angina (2), deafness (2) and diabetes (2). Coroners briefs were available for 10 of the 15 drivers aged 60 or over. Reading of the briefs showed that the cause of death was related to heart disease, rather than CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 13

injuries sustained in the crash, for seven of these ten drivers. Other diseases were noted as possible causes of the crash for two other drivers. For only one driver was it noted that death resulted from injuries sustained in the crash. 4.3 LICENSING The types of licences held by the drivers and riders are summarised in Table 4.6. Overall, 13% of drivers or riders did not hold a current licence. Of these, 6% had never held a licence (all aged under 25), 6% had their licence disqualified and 2% had allowed their licence to lapse. Table 4.6. Type of licence held by the driver or rider. Some drivers or riders held multiple types of licence. Type of licence held Number of crashes Percent of crashes Never held licence 7 6 Licence lapsed 2 2 Licence disqualified 7 6 Learner 5 4 Probationary 25 20 Full car 73 58 Motorcycle 9 7 Light truck 3 2 Heavy truck/bus 6 5 Articulated truck 7 6 Other 3 2 Unknown 2 2 Five car drivers (4%) held a learners permit. At the time of the crash two of the learners were not accompanied by a licensed driver (there were no passengers). Of the 25 probationary drivers, only two were aged 25 and over. Seven of the ten motorcycle riders in the crashes held a current motorcycle licence. One rider had never been licensed, one was disqualified and information is unavailable for the remaining rider. Information about driver training was available for 101 drivers or riders. Five car drivers had undertaken a training course of some kind. One motorcyclist had completed a course for beginners. 14 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

4.4 ALCOHOL AND OTHER DRUGS There were several sources of information about the presence of alcohol and other drugs in crashed drivers. Toxicology reports were prepared for deceased drivers. Some injured drivers had blood tests taken for analysis for alcohol and/or other drugs. Some uninjured drivers were breath tested for alcohol. 4.4.1 Toxicology reports Toxicology reports were available for 98 drivers or riders and 24 passengers from 118 crashes. These reports are summarised in Table 4.7. Forty-two drivers or riders (43% of reports) were drug free. The general BAC limit of.050 was exceeded by 36 deceased drivers or riders (37% of reports), but in 22 of these cases, other drugs were also present. Cannabinoids were detected in 22 drivers or riders (22% of reports), in all but four cases this was in conjunction with alcohol. This pattern is shown in Figure 4.2. Figure 4.2. The prevalence of alcohol, cannabis and other drugs in the 98 deceased drivers and riders for whom toxicology reports were available. In this figure, persons with BAC less than.050 and no other drug are classified drug-free. Other drugs 14% Alcohol and other drugs 4% Alcohol 14% Alcohol and cannabis 18% No drugs 46% Cannabis 4% CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 15

Table 4.7. Summary of toxicology reports received. Type of drug identified Driver or rider Passenger Number Percent Number Percent No drugs identified 42 43 10 42 Alcohol (no other drugs) 0.001% to 0.050% 2 2 1 4 0.051% to 0.149% 3 3 4 17 >=0.150% 11 11 2 8 Cannabinoids (no alcohol) 4 4 1 4 Opiates 1 1 0 0 Benzodiazepines 3 3 0 0 Stimulants 1 1 0 0 Antidepressants 1 1 0 0 Analgesics 4 4 0 0 Alcohol and cannabinoids... 0.001%-0.050% 0 0 2 8 0.051%-0.149% 3 3 1 4 >=0.150% 13 13 1 4 Alcohol >=0.150%, cannabinoids and benzodiazepines Alcohol >=0.150%, cannabinoids, opiates and benzodiazepines 1 1 0 0 1 1 2 8 Alcohol and other drugs (not cannabinoids) 0.001%-0.050% 0 0 0 0 0.051%-0.149% 1 1 0 0 >=0.150% 3 3 0 0 Opiates and benzodiazepines 2 2 0 0 Other drugs or combinations 2 2 0 0 Total persons 98 99 24 99 Note: Detection of cannabinoids cannot be taken to imply that the driver/rider was affected by this drug at the time of the crash. Detection implies only that the substance was used during the last several weeks. 4.4.2 Alcohol and cannabis Table 4.8 summarises the availability of data relating to alcohol and other drugs. For the 24 surviving drivers, results of breath tests for alcohol were available for nine drivers and results of blood tests for alcohol were available for 14 drivers. Toxicology reports were received for 98 of the 103 deceased drivers. 16 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

Table 4.8. Summary of information about alcohol and other drugs in crashed drivers. Information available Surviving drivers (n=24) Deceased drivers (n=103) Breath test for alcohol 9 0 Blood test for alcohol 14 0 Toxicology report 0 98 In deceased drivers and riders, cannabis was coded as present if the toxicology report stated that the metabolite (as carboxy-thc) was found in blood or urine. This could result in drugs being coded as present which may have been taken more than 12 hours (or even several weeks) before the crash. Only eight of the 24 surviving drivers and riders provided information about cannabis use in the 12 hours prior to the crash. Each of these drivers said that they had not used cannabis during that period. The reliability of those responses was not able to be tested. The concentrations of the cannabis metabolite detected in killed drivers are summarised in Appendix 4. Table 4.9 summarises the available BAC and cannabis information. Overall, BAC was unavailable for about 5% of crashes and cannabis unavailable for 17% of crashes. BACs of greater than.05 were recorded for 35% of crashes (36% of known). BACs of greater than.15 were found in 26% of crashes (27% of known). Cannabis and alcohol were often found together. For BACs over.05, cannabis was present in 41% (49% of known). For BACs over.15, cannabis was present in 45% (52% of known). There were only four cases of cannabis (4% of known) where BAC<.05. Table 4.9. Numbers of drivers and riders with given blood alcohol concentrations (BACs) with and without cannabis present. BAC Without cannabis (n=83) With cannabis (n=22) Cannabis unknown (n=22) Total (n=127) <=.050 64 4 9 77.051-.149 5 3 3 11 >=.150 14 15 4 33 Unknown 0 0 6 6 4.4.3 Correlates of alcohol use 4.4.3.1 Sex of driver or rider Overall, 89% of the drivers or riders with BAC>.05 in the study were male. Forty percent of males had BAC>.05, compared with 22% of females (see Table 4.10). CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 17

Table 4.10. Percentages of drivers and riders with given blood alcohol concentrations (BACs) as a function of driver or rider sex. BAC Sex of driver or rider Male (n=102) Female (n=25) Total (n=127) <=.050 58 72 61.051-.149 10 4 9 >=.150 28 16 26 Unknown 4 8 5 Total 100 100 101 The proportions of drink drivers whose BAC was over.15 were similar for males and females. While females in single vehicle crashes were less likely to have been drink driving than males, those who had been drink driving were just as likely as males to have had a very high BAC. 4.4.3.2 Age The percentages of drivers and riders of each age group at each BAC level are presented in Table 4.11. None of the drivers or riders aged 60 or over had BACs greater than.04. The proportion of drivers or riders with BAC>.05 was highest for under 25s but those aged 25 to 59 were more likely to have BACs exceeding.15. Table 4.11. Percentages of drivers and riders with given blood alcohol concentrations (BACs) as a function of age group. BAC Age group Under 25 (n=52) 25 to 59 (n=60) 60 and over (n=15) Total (n=127) <=.050 56 60 80 61.051-.149 17 3 0 9 >=.150 25 33 0 26 Unknown 2 3 20 5 Total 100 99 100 101 4.4.3.3 Level of education The relationship between alcohol and level of education is complicated by the relationship between age and level of education. Table 4.12 presents the percentages of drivers or riders with BAC>.05 as a function of level of education and age group. For under 25 year olds, level of education did not affect the percentage with BAC>.05. For 25 to 59 year olds, however, none of those with TAFE or university education had BAC>.05. 18 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

Table 4.12. Percentages of drivers and riders who had BAC>.05 for each highest level of education received and age group. Highest level of Age group education Under 25 (n=52) 25 to 59 (n=60) 60 and over (n=15) Primary or secondary 41 44 0 TAFE or university 43 0 0 4.4.3.4 Location of crash Similar percentages of crashes in the metropolitan area (36%) and the rest of the study area (33%) involved BACs>.05. Drink driving crashes were more common on major roads (59% of crashes on these roads) than on freeways (29%), highways (29%) or minor roads (27%). 4.4.3.5 Licensing Of the 44 drivers or riders with BACs>.05, 22 had full car licences and two riders had motorcycle licences. Eleven of the 44 were novice drivers: nine had probationary licences (3 of these were BAC>.15) and two had learner permits. Five drink drivers had never held a licence, three were disqualified (including one motorcyclist) and one licence had lapsed. One learner driver and two probationary drivers had BACs between zero and.05, in contravention of the licensing restrictions. 4.4.4 Correlates of cannabis use 4.4.4.1 Sex of driver or rider Cannabis was detected for 20% of male drivers and 8% of female drivers. Overall, 91% of the drivers for whom cannabis was detected were male. 4.4.4.2 Age Cannabis was detected in 17% of drivers or riders aged under 25 and 22% of those aged 25 to 59. It was not detected in any drivers or riders aged 60 or over. 4.4.4.3 Level of education Table 4.13 shows that the proportion of drivers aged under 25 in whom cannabis was detected did not differ according to level of education but that drivers and riders aged 25 to 59 with TAFE or university education were less likely to have used cannabis. This was the same pattern found for BAC>.05. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 19

Table 4.13. Percentages of drivers and riders with cannabis detected for each highest level of education received and age group. Highest level of Age group education Under 25 (n=52) 25 to 59 (n=60) 60 and over (n=15) Primary or secondary 18 24 0 TAFE or university 14 0 0 4.4.4.4 Location of crash Similar percentages of crashes in the metropolitan area (16%) and the rest of the study area involved use of cannabis (20%). Cannabis-involved crashes were less common on highways (10%) than freeways (21%), major roads (24%) and minor roads (23%). 4.4.4.5 Licensing Of the 22 drivers and riders in whom cannabis was detected, 12 had a full car or motorcycle licence. Five were novice drivers: one held a learner s permit and four had probationary licences. Five drivers or riders were unlicensed: two had never had a licence, two had their licences disqualified and one had a licence which had lapsed. 4.4.5 Prescription and nonprescription drugs Surviving drivers in crashes were asked if they had taken any prescription drugs or nonprescription drugs in the 12 hours prior to the crash. For deceased drivers this information was requested from relatives or treating doctors. Non-prescription drugs were described as medications that you can buy only from the chemist, but without a prescription from a doctor. Information about consumption of prescription and nonprescription drugs was available for 99 and 93 cases, respectively. Seventeen drivers or riders (17% of known) had taken prescription drugs in the 12 hours prior to the crash. This included 14% of under 25s, 18% of 25 to 59s and 27% of those aged 60 and over. Five drivers with BAC>.15 and one driver with BAC>.05 had taken prescription drugs in the previous 12 hours. The prescription drugs which had been taken included asthma preparations (3 persons), benzodiazepines (3), antidepressants (2), blood pressure medication (2), insulin and heart medication. Only three drivers or riders had taken nonprescription drugs in the 12 hours prior to the crash. Two were aged under 25 and one was aged 25 to 59. One of the drivers who had taken nonprescription drugs had BAC>.15. 20 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

4.5 PASSENGERS There were 74 passengers in 48 crashes (38%). The number of passengers was unknown for one crash. The number of passengers per vehicle ranged from 0 to 5. Two of the passengers were pillions on motorcycles (both killed). In total, 30 passengers were killed in the crashes. As Table 4.14 shows, passengers were more likely to be of the same sex as the driver than to be of the opposite sex or at least one of each sex. The young drivers were more likely to have passengers than the older drivers: 56% under 25 drivers had one or more passengers, 25% of 25-59 year old drivers had passengers, 27% of 60 and over drivers had passengers. Table 4.14. Percentages of crashes according to the sex of driver and sex of passengers. Crashes with passengers only. Sex of Sex of driver passengers Male (n=37) Female (n=11) Total (n=48) Male 59 27 52 Female 24 45 29 Mixed 16 27 19 Total 99 99 100 Table 4.15 shows that most passengers were in the same age group as the driver of the vehicle. This was especially true for under 25 year olds. Table 4.15. Percentages of crashes according to the age of driver and age of passengers. Crashes with passengers only. The percentages may not add to 100% because there may be passengers in more than one age group in a single crash. Age of passengers Age of driver Under 25 (n=29) 25 to 59 (n=15) 60 and over (n=4) Overall (n=48) 0-14 0 33 0 10 15-24 90 27 0 63 25-59 14 66 50 33 60 and over 0 13 50 8 The BAC of the passenger(s) was known for 24 crashes. In four of these crashes, the BAC of the driver was greater than that of the passenger. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 21

22 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

5. VEHICLES INVOLVED A number of characteristics of the vehicles involved in crashes were recorded and are summarised here. It should be noted that many of the vehicle characteristics may not have contributed directly to the occurrence or severity of the crashes, but rather may reflect characteristics of the drivers or riders. This section presents the characteristics of the vehicles involved in the fatal single vehicle crashes. Comparisons between these vehicles and a control sample of vehicle are presented in the companion report, Estimation of risk factors for fatal singlevehicle crashes. 5.1 TYPES OF VEHICLES INVOLVED Almost 85% of the fatal single vehicle crashes involved a car (see Table 5.1). Motorcycles comprised about 8% and light commercial vehicles about 6% of vehicles in the study. The remaining 2% of vehicles were trucks. Table 5.1. Types of vehicles involved in the fatal single vehicle crashes. Type of vehicle Number of crashes Percent of crashes Car 107 84 Motorcycle 10 8 Light commercial vehicle 7 6 Truck 3 2 Total 127 100 Engine capacity was known for nine of the ten motorcycles in the study. Six of the nine motorcycles had an engine capacity of over 500 cc. None of the motorcycles with an engine capacity greater than 260 cc was ridden by a learner or probationary rider. The years of manufacture of the vehicles involved in the crashes ranged from 1956 to 1996 (see Table 5.2). Almost 20% of the vehicles involved in crashes were manufactured before 1978 and so were not required to comply with some of the current safety standards. The proportions of drivers who drove vehicles manufactured before 1978 appeared to increase with age but this effect was not statistically significant (13% under 25s, 21% 25 to 59s, 33% 60 and over, χ 2 (2)=3.3, p>.10). Seven vehicles (6%) were not currently registered. Four of these vehicles had drivers with BAC>.05 (one of whom was disqualified and one who had never obtained a licence). Another vehicle was driven by a driver whose licence was lapsed and another by a driver who had never obtained a licence. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 23

Table 5.2. Years of manufacture of vehicles involved in the fatal single vehicle crashes. Year of manufacture Number of crashes Percent of crashes Pre-1970 2 2 1970-1977 20 16 1978-1989 61 48 1990 and on 32 25 Unknown 12 9 Total 127 100 The most common makes of vehicles involved in the crashes were: Holden (31), Ford (25), Toyota (15) and Mitsubishi (11). Model information was not available for all vehicles but the most common models reported were Holden Commodore (17) and Ford Falcon/Fairmont/Fairlane (13). Nine vehicles were recorded as having been modified. For two vehicles, the modification was conversion to LPG. No vehicles were coded as having had engine or driveline modifications or structural modifications. Anti-lock brakes were recorded as present for seven vehicles (6%) and driver s side airbags for six vehicles (5%). Cruise control was fitted to six vehicles but was not in use at the time of the crash for the four vehicles for which this information was available. 5.2 HEATING AND VENTILATION Airconditioning was fitted to 51 vehicles (40%) and turned on for 2 of the 35 vehicles for which this information was available. Both of these vehicles crashed during daytime. Whether the heater was on at the time of the crash was known for 83 of the 127 vehicles. The heater control was in the on or midway position for 43 vehicles. For those crashes where it was known, the heater was on for 42% of daytime crashes and 49% of night-time crashes. Whether the ventilation controls were pointing to recirculate or fresh air was recorded for 57 vehicles. For 20 vehicles (35%), the air was being recirculated. Recirculated air was more common in vehicles which crashed at night than those which crashed during the day (13/31 versus 7/26). Whether any windows were open at the time of the crash was known for 78 vehicles. For 18 (23%) of these vehicles, at least one window was open. The window was open for 6/31 daytime crashes compared with 12/47 night-time crashes. 24 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

5.3 SOUND Whether the radio, cassette, CD or CB radio was turned on at the time of the crash was known for 63 crashes. Of these crashes, it was turned on for 35 cases (66%). The sound was more likely to be turned on in vehicles which crashed at night than during the day. 5.4 VEHICLE DEFECTS Vehicle defects were not recorded on the questionnaire but a small number were mentioned in the Coronial briefs. They included: extensive rust which was considered to have weakened the car, brake defects which may have played a role in the crash, and rust and insufficient tread. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 25

26 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

6. LOCATIONS OF CRASHES More crashes occurred in the Melbourne metropolitan area (60%) than in the rest of the study area (40%). For those crashes in the metropolitan area, the land use adjacent to the crash site was most commonly urban residential (see Table 6.1). For those crashes in the rest of the study area, the most common adjacent land use was rural farmland. Table 6.1. Percentages of crashes according to land use adjacent to the crash site. Adjacent land use Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) Urban residential 47 10 32 Urban commercial 13 0 8 Urban industrial 7 0 4 Urban parkland 8 0 5 Rural residential 8 14 10 Rural farmland 16 59 33 Rural forest 1 16 7 Unknown 0 2 1 Total 100 101 100 6.1 TYPE OF ROAD Table 6.2 summarises the type of road on which the crash occurred. The categories were based on Melway Directory markings for the Melbourne metropolitan area and the VicRoads Country Directory markings for the rest of the study area. Most crashes occurred on highways, with fewer crashes occurring on major roads and minor roads. When a crash occurred at an intersection, the road type was coded with respect to the first-mentioned road (it was unclear from some notifications the road on which the vehicle was travelling). Table 6.2. Percentages of crashes which occurred on each type of road. Road type Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) Freeway 9 14 11 Highway 49 31 42 Major road 24 22 23 Minor road 17 33 24 Off-road 1 0 1 Total 100 100 101 A larger proportion of crashes occurred on highways and a smaller proportion on minor roads in the metropolitan area compared to the rest of the study area (see Table 6.2). CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 27

Overall, 66% of the crashes occurred on undivided roads (58% of metropolitan, 77% of rural). Most crashes in both areas occurred on two way undivided roads, although a somewhat larger proportion of metropolitan than rural crashes occurred on dual carriageways without service roads (see Table 6.3). Table 6.3. Percentages of crashes according to road configuration. Road configuration Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) One way 0 0 0 Two way undivided 57 75 64 Two way undivided with 1 2 2 service roads Dual carriageway without 33 20 28 service roads Dual carriageway with 7 2 5 service roads Other 3 0 2 Unknown 0 2 1 Total 101 101 102 6.2 SPEED LIMIT Of the crashes in the study, 48% occurred where the speed limit was 100 km/h or greater (see Table 6.4). Not surprisingly, a higher proportion of rural crashes (82%) than metropolitan crashes (23%) occurred in these higher speed zones. Table 6.4. Percentages of crashes according to speed limit at the crash site. Speed limit (km/h)* Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) 40 3 0 2 60 41 6 27 70 13 2 9 80 16 6 12 90 3 0 2 100 22 73 43 110 1 9 5 Unknown 1 4 2 Total 100 100 102 * Includes temporary speed limits because of roadworks. 6.3 ROAD GEOMETRY Information about horizontal alignment was available for 119 crash sites. Where the recorded radius of the curve was greater than 1000m, the road was deemed to be straight. Where it was known, 32% of metropolitan crashes and 40% of rural crashes were on curves. Table 6.5 shows that the most common curve radii were 100 to 200m and between 600m and 100m. 28 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

Table 6.5. Curve radius for crashes on curves. Curve radius (m) Number of cases Percent of cases Less than 100 5 12 100 to 200 10 24 200 to 300 1 2 300 to 400 4 10 400 to 500 3 7 500 to 600 0 0 600 to 1000 8 19 Unknown 11 26 Total 42 100 6.3.1 Type of intersection Overall, 20% of the crashes in the study occurred at intersections. These comprised 29% of crashes in the metropolitan area and 8% of rural crashes. Table 6.6 shows that T-intersections were the most common types of intersections for both metropolitan and rural crashes. In almost all of these crashes, the vehicle was travelling on the continuing leg of the T-intersection. Table 6.6. Percentages of intersection crashes which occurred at each type of intersection. Type of intersection Metropolitan area (n=22) Rest of study area (n=4) Overall (n=26) X 23 25 23 Y 17 0 15 T 55 50 54 Multi 0 0 0 Unknown 5 25 8 Total 100 100 100 6.4 DELINEATION For about half of the crashes, there was one lane only in the direction of travel. This was more common outside the metropolitan area (see Table 6.7). Lane configuration at the site of the crash was recorded. Table 6.8 shows that the number of lanes was not changing at most crash sites. Exit lanes were the most common feature where the number of lanes was changing in the metropolitan area. Road markings were present at about 90% of the crash sites (91% of metropolitan, 88% of rural). There were raised reflective pavement markers at 22% of the crash sites (18% metropolitan, 28% rural). CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 29

Table 6.7. Percentages of crashes according to the number of lanes in the direction of travel. Number of lanes Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) 1 43 67 53 2 36 27 32 3 14 2 9 4 4 0 2 Unknown 3 4 3 Total 100 100 99 Table 6.8. Percentages of crashes according to lane configuration. There can be more than one per crash, so the percentages sum to more than 100%. Lane configuration Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) Number of lanes not 87 94 90 changing Breakdown lane 3 0 2 Climbing lane 1 2 2 Exit lane 8 0 5 Entry lane 3 0 2 Turn slot 4 0 2 Bus/transit lane 1 0 1 Tram lines/safety zones 4 0 2 6.5 ROAD FURNITURE There was a median present at the site of 39% of crashes (46% of metropolitan, 29% of rest of study area). Table 6.9 shows that most medians comprised paved, gravel or low level landscape (with or without trees). Table 6.9. Percentages of crashes which occurred at sites with each type of median. Type of median Metropolitan (n=35) Rest of study area (n=15) Overall (n=50) Painted - wider than 2 lines 3 13 6 Raised studs 3 0 2 Paved, gravel or low level 26 13 22 landscape Paved, gravel or low level 43 47 44 landscape with trees Physical barrier 14 7 12 Tram safety zone 6 0 4 Other 3 13 6 Unknown 3 7 4 Total 101 100 100 30 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

For the seven sites where there were service lanes, the type of outer separator was coded. The outer separators were all coded as being paved or gravel or low level landscaping. About half of them had trees. There was a side drain present for 57 crashes (45%). Information about the presence of a side drain was absent for 2 crashes (2%). The side drain was considered to be navigable in 39 cases. Kerbs were present at the sites of 44% of crashes (68% of metropolitan, 8% of rural). Approximately half of the kerbs were mountable and half were barrier kerbs. 6.6 TRAFFIC CONTROLS AND LOCAL AREA TRAFFIC MANAGEMENT DEVICES Traffic controls were present at the site of 13% of crashes (18% of metropolitan, 6% of rural). Traffic control signals were the most common form of traffic control, followed by Stop signs and roundabouts (see Table 6.10). Table 6.10. Percentages of crash sites by type of traffic control present - sites with traffic controls present only. Traffic controls present Metropolitan area (n=14) Rest of study area (n=3) Overall (n=17) Stop sign 21 33 24 Give Way sign 7 0 6 Roundabout 21 0 18 Double unbroken lines 8 0 6 Traffic control signals 36 0 29 Pedestrian crossing 7 0 6 Warning signs 0 33 6 Unknown 0 33 6 Total 100 99 101 Local area traffic management devices were present at eight crash sites (6%), all of which were in the metropolitan area. Five of the eight LATMs were roundabouts. 6.7 THE ROAD SURFACE Only 3% of crashes occurred on unsealed roads (see Table 6.11). Most road surfaces were coded as asphalt. Chip seal was somewhat more common and slick bitumen less common outside of the metropolitan area. The surface coarseness of the road pavement was coded as medium for most crashes. Coarse pavements were somewhat more common outside the metropolitan area (see Table 6.12). CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 31

Table 6.11. Percentages of crashes according to the road surface. Road surface Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) Gravel/crushed rock 1 6 3 Chip seal 9 25 16 Asphalt 65 53 60 Slick bitumen 24 12 19 Other 1 2 2 Unknown 0 2 1 Total 100 100 101 Table 6.12. Percentages of crashes according to the coarseness of the road surface. Surface coarseness Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) Coarse 5 18 10 Medium 80 73 78 Fine 13 6 10 Unknown 1 4 2 Total 99 101 100 Mud, oil or loose material on the road surface was present for only a small number of crashes overall, but was present for 10% of rural crashes (see Table 6.13). Table 6.13. Percentages of crashes according to the cleanness of the road surface. Surface cleanness Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) Clean 97 88 94 Muddy 1 0 1 Oily 0 0 0 Loose material 1 10 5 Unknown 0 2 1 Total 99 100 101 Surface irregularities such as pot holes, pit lids and drainage grates, deformed pavement and sudden changes in the road surface were reported for very few crashes. 6.7.1 Road shoulder The road had a shoulder for 62% of the metropolitan crash sites and 96% of the rural crash sites. At metropolitan crash sites, 49% of shoulders were sealed compared with 20% of shoulders at rural crash sites (see Table 6.14). 32 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

Table 6.14. Percentages of crashes according to the material of the shoulder - crashes where there was a shoulder present only. Surface of shoulder Metropolitan area (n=47) Rest of study area (n=49) Overall (n=96) Gravel/crushed rock 40 63 52 Chip seal 9 2 5 Asphalt 32 16 24 Slick bitumen 9 2 5 Other 11 16 14 Total 101 99 100 6.8 ROADSIDE OBJECTS The presence of roadside objects was coded for each crash location. Trees were the most common roadside object in the metropolitan area and the rest of the study area. This was followed by poles and fences in the metropolitan area and fences and poles in the rest of the study area (see Table 6.15). Table 6.15. Percentages of crashes according to roadside objects present. More than one roadside object may be coded at a site. Roadside objects present Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) Trees 80 84 82 Poles 71 35 57 Fences 57 43 51 Buildings 33 4 21 Guard rails 11 4 8 Railway crossing 0 0 0 Bridge abutment 3 2 2 Culvert end walls 4 2 3 Other 11 14 12 6.9 ENVIRONMENTAL FACTORS The road was wet for 37 crashes (29%). Half of the crashes occurred when it was dark (see Table 6.16). There were no street lights at the site of 32% of the crashes in the metropolitan area and 86% of crashes in the rest of the study area. Glare from oncoming headlights was considered to be a potential problem in 5 cases (4%). In 2 cases glare from the sun may have reduced visibility. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 33

Table 6.16. Light conditions at the time of the crash. Light conditions Number of crashes Percent of crashes Percent of daytime crashes Sunny/bright sunlight 25 20 39 Overcast/cloudy 30 24 47 Dusk or dawn 7 6 11 Dark 63 50 Unknown 2 2 3 Total 127 100 34 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

7. CRASHES INVOLVING TREES AND POLES Almost 75% of the crashes involved an impact with a tree or pole or both (71% of metropolitan crashes, 78% of rural crashes). Trees and poles were involved to the same extent in metropolitan crashes but trees were much more common in rural crashes (see Table 7.1). Overall, 53% of the trees and 36% of the poles impacted were on the right of the direction of travel of the vehicle. Table 7.1. Percentages of crashes which involved an impact with a tree or a pole. Object impacted Metropolitan area (n=76) Rest of study area (n=51) Overall (n=127) Tree 33 65 46 Pole 34 12 25 Tree and pole 4 2 3 No tree or pole impacted 29 22 26 Total 100 101 100 Impacts with trees or poles did not appear to be related to driver age group or BAC level (see Table 7.2). Table 7.2. Percentages of crashes which involved an impact with a tree or a pole according to the age group and BAC of the driver. Driver Percent hit trees Percent hit poles characteristics Metropolitan Overall Metropolitan area area Rest of study area Rest of study area Overall Age Under 25 58 43 53 24 43 30 25 to 59 56 48 51 28 30 29 60 and over* 27 50 33 27 25 27 BAC<=.05 48 52 49 29 33 30 BAC>.05 60 44 54 24 38 29 * percentages may not be reliable because of small numbers in this group - 11 metropolitan crashes and 4 crashes in the rest of the study area 7.1 DISTANCES TO IMPACTED TREES AND POLES The distances to the trees and poles which vehicles impacted were measured. For trees or poles on the same side of the road as the vehicle was travelling (hereafter referred to as on the left), the distance from the edgeline of the road to the tree or pole was measured (perpendicular to the road). For trees or poles on the opposite side of the road (hereafter referred to as on the right), the distance from the centreline of the road to the tree or pole was measured (perpendicular to the road) and also the distance from the right edgeline to the tree or pole. For trees and poles in central medians, the distance measured was from the edge of the rightmost travel lane to the tree or pole. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 35

Figure 7.1 summarises the distances from the closest edgeline to impacted trees and poles. It shows that many of the trees impacted were between 3m and 10m from the travel lane. Eleven trees and twelve poles were less than 3m from the edgeline. Seven trees and nine poles were less than 2m from the edgeline. Figure 7.1. Distances to impacted trees and poles. 22 20 Number of crashes 18 16 14 12 10 8 6 4 2 Tree Pole 0 Less than 2m 2m to <3m 3m to <5m 5m to <10m 10m and over Unknown Distance from edgeline The figures below show the distances to trees and poles impacted in the metropolitan area and the rest of the study area. In the metropolitan area, crashes involving poles were more frequent (see Figure 7.2). Poles were often less than 2m from the edge of the road. The distances to trees varied. In the rest of the study area, many more crashes involved trees than poles (see Figure 7.3). Most trees were within 3 to 10m of the edgeline. In 65% of crashes where a tree was impacted, it was on the right. 36 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

Figure 7.2. Distances to trees and poles impacted in the metropolitan area. 16 14 Number of crashes 12 10 8 6 4 2 Tree Pole 0 Less than 2m 2m to <3m 3m to <5m 5m to <10m 10m and over Unknown Distance from edgeline Figure 7.3. Distances to trees and poles impacted in the rest of the study area. 16 14 Number of crashes 12 10 8 6 4 2 Tree Pole 0 Less than 2m 2m to <3m 3m to <5m 5m to <10m 10m and over Unknown Distance from edgeline CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 37

7.2 COMPARISONS WITH VICROADS CLEAR ZONE POLICY The VicRoads clear zone policy (Road Design Note 3-1a, March 1995) outlines desirable widths for clear zones according to speeds, traffic volumes and road curvature. The minimum desirable clear zone width is 3m and this increases to 9m for 85th percentile speeds of 95 km/h and over with approach volumes of greater than or equal to 5,000 vehicles/day. Application of the clear zone policy to the crash data was complicated by the lack of traffic volumes for the crash sites and the lack of 85th percentile speeds. A conservative comparison was undertaken, which assumed the lowest traffic volume and that the 85th percentile speed was the speed limit. A summary of this comparison is presented in Table 7.3. It shows that 22 trees and 14 poles were located within the desirable clear zone. The numbers in the table are underestimates because they do not include the 13 trees and 17 poles for which the distances were unknown. Table 7.3. Numbers of crashes involving trees and poles within the desirable clear zone for that speed. The distances are from the edgeline of the road closest to the tree or pole. Speed limit (km/h) Clear zone width Impacted trees within clear zone Impacted poles within clear zone 60 3.0 m 6 5 70 3.4 m 0 4 80 4.2 m 1 1 100 6.0 m 14 3 110 6.0 m 1 1 Total 22 14 The clear zone policy notes that in urban areas it is often not practical to provide even three metres because of poles and trees. In these circumstances, a clear zone of one metre may have to be adopted. Any clearances less than one metre could adversely affect both the safety and operating conditions of traffic (p.2). In the sample of crashes analysed here, one tree and four poles were less than one metre from the edge of the traffic lane. 7.3 CRASHES WITH TREES AND POLES ON CURVES Of the 94 crashes involving trees and poles, 25 (27%) occurred on a curve. For comparison, 36/118 (31%) crashes in total occurred on a curve. The direction of curvature of the road and the direction to the pole or tree that was impacted are compared in Table 7.4. For the 24 crashes where data was available, in 16 crashes the tree or pole was on the outside of the curve. Distance to the pole or tree was known for 12 of the 16 crashes where the tree or pole was on the outside of the curve. For 5 of these 12 crashes, the tree or pole was within the clear zone for that speed limit even if the road was straight. For four of the seven crashes where the tree or pole was outside the clear zone for a straight section of road, 38 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

it was within the clear zone calculated according to the radius of curvature and speed limit. Thus, a total of 40 crashes (31%) in the study occurred when the tree or pole impacted was within the recommended clear zone. Table 7.4. Numbers of crashes on curves according to the direction of curvature of the road and the direction to the pole or tree that was impacted. Direction of curvature Direction of pole or tree Left (n=10) Right (n=14) Unknown (n=1) Left (n=14) 4 10 0 Right (n=10) 6 4 0 Unknown (n=1) 0 0 1 CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 39

40 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

8. SUMMARY OF CHARACTERISTICS OF CRASHES There were 127 single vehicle crashes in which 133 persons were killed. The crashes had the following characteristics: Type of crash almost 75% of crashes involved an impact with a tree or pole or both (71% metropolitan, 78% of rest of study area crashes) tree crashes were more common outside the metropolitan area 53% of trees and 36% of poles impacted were on the right-hand side of the road 31% of crashes involved impact with a tree or pole located within the desirable clear zone outlined in VicRoads policy Temporal characteristics were most common from midnight to 6 am and midday to 6 pm and least frequent from 6 am to midday were most common on Saturdays and Sundays were more common in February, April and September than in other months Persons killed five crashes were multiple fatalities the driver was killed in 81% of crashes almost half the persons killed were aged under 25 Drivers and riders 80% of the drivers and riders were male more than 40% of the drivers and riders were aged under 25 natural disease (particularly heart disease) was the cause of death for at least 60% of the drivers aged 60 and over 13% of the drivers had been severely depressed in the previous six months 13% of drivers or riders did not hold a current licence (6% never held a licence, 2% licence lapsed, 6% licence disqualified) Alcohol and other drugs BAC was known for 95% of crashes and cannabis use was known for 83% of crashes. The percentages in this section are percentages of known cases. All of the cases of cannabis were in deceased drivers where it was detected as the metabolite, carboxy-thc. This substance may be detected some weeks after cannabis use and therefore its detection does not prove that the driver was under the influence of cannabis at the time of the crash. BAC>.05 in 36%, BAC>.15 in 27% of crashes cannabis was most commonly found in conjunction with alcohol: for BAC>.05, cannabis was present in 49% for BAC>.15, cannabis was present in 52% but cannabis was present in only 6% for BAC<.05 males were twice as likely as females to have BAC>.05 but the proportions of these groups with BAC>.15 were similar CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 41

no alcohol or cannabis was identified in drivers aged 60 and over drivers aged 25 to 59 were more likely to have BAC>.15 than those under 25 higher levels of education were related to fewer drivers with BAC>.05 or cannabis for 25 to 59 year olds but not under 25s proportions of crashes with BAC>.05 or cannabis use were similar in metropolitan and rural areas BAC>.05 more often on major roads than freeways, highways or minor roads cannabis was detected for 23% of male drivers and 11% of female drivers cannabis was detected in 21% of those under 25 and 25% of those aged 25 to 59 cannabis-involved crashes were less common on highways than freeways, major roads or minor roads information about consumption of prescription and nonprescription drugs was available for 78% and 73% of drivers, respectively 17% of drivers had taken prescription drugs in the 12 hours prior to the crash (14% of under 25s, 18% of 25 to 59s and 27% of those aged 60 and over) 3 drivers or riders had taken nonprescription drugs in the 12 hours prior to the crash Passengers there were 74 passengers of whom 30 were killed passengers were more likely to be of the same sex as the driver than to be of the opposite sex or a mixture of males and females young drivers were more likely to have passengers than other drivers most passengers were in the same age group as the driver of the vehicle (particularly for under 25s) in four of the 24 crashes for which BAC of both driver and (killed) passenger were known, the BAC of the driver was greater than that of the passenger. Vehicles 10 crashes involved motorcycles, 7 involved light commercial vehicles and 3 involved trucks six of the 10 motorcycles had an engine capacity of greater than 500 cc but none of the motorcycles with an engine capacity greater than 260 cc was ridden by a learner or probationary rider almost 20% of the vehicles involved in crashes were manufactured before 1978 and so were not required to comply with some of the current safety standards age of the vehicle was not related to age of the driver Locations of crashes 60% of crashes occurred in Melbourne metropolitan area, 40% in the remainder of the study region most crashes occurred on highways (42%) a larger proportion of crashes occurred on highways and a smaller proportion on minor roads in the metropolitan area compared to the rest of the study area two-thirds of the crashes occurred on undivided roads (58% of metropolitan, 77% of rest of study area) half of the crashes occurred where the speed limit was 100 km/h or 110 km/h (23% of metropolitan, 82% of rest of study area) 42 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

about one-third of the crashes occurred on curves (27% of metropolitan, 36% of rest of study area) 20% of metropolitan and 8% of rest of study area crashes occurred at intersections (most commonly T intersections) traffic controls were present at the site of 13% of crashes (18% of metropolitan, 6% of rest of study area) mud, oil or loose material was present on the road surface for 10% of crashes in the rest of the study area the road had a shoulder for 62% of metropolitan crashes and 96% of rest of study area crashes. In the metropolitan area, 49% of shoulders were sealed compared with 20% with shoulders in the rest of the study area CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 43

44 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

9. DISCUSSION The purpose of this report was to present a more detailed description of the characteristics of fatal single vehicle crashes than is currently available. It complements the Coronial findings which will be made for most of the crashes in the sample. While this report provides those detailed characteristics of crashes, it does not include comparisons with controls who have not crashed or comparisons with other types or severities of crashes. Thus the scope for identification of risk factors and possible countermeasures is limited. Nevertheless, the study suggests that the main identifiable factor contributing to the occurrence of fatal single vehicle crashes was BAC>.05 (and particularly BAC>.15) and the main identifiable factors contributing to the severity of the crashes appeared to be: impacts with trees and poles and some older vehicles which are unlikely to have complied with most of the current safety standards. 9.1 ALCOHOL AND CANNABIS Many of the findings summarised in Chapter 8 are associated with drink driving. For example, the time of day and day of week patterns found reflect drink driving patterns. When only crashes with BAC<.05 are examined, the timing of crashes becomes much more uniform. Unlicensed driving was also most commonly found in conjunction with alcohol. In very few cases was cannabis found in the absence of alcohol. Cannabis was most commonly associated with high levels of alcohol and the behavioural effect of the alcohol is likely to have predominated over that of cannabis. The lack of a significant number of crashes where cannabis alone was present is in agreement with recent studies that suggest that the behavioural effects of cannabis alone may not contribute greatly to crash causation (Drummer, 1995). Cannabis was most commonly detected at autopsy and it was the metabolite, carboxy- THC that was detected in blood or urine. Given the long half-life of this substance, it is possible that some of the drivers or riders for whom cannabis was detected may not have been behaviourally affected by it at the time of the crash. The results of the current study are compared with those of two other recent Australian studies in Table 9.1. The prevalence of alcohol only was somewhat lower in the current study than in the other studies. There are two likely reasons for this. First, alcohol is probably less common in drivers in Victoria than in South Australia (Longo, Hunter and Lokan, 1996) or in New South Wales and Western Australia (Drummer, 1994). Indeed, the overall prevalence of alcohol in crashes (single vehicle and multi-vehicle) in Drummer s study ranged from 32% in Victoria to 44% in Western Australia. Secondly, there may have been a decrease in drink driving from the time of the earlier studies to the time of the current study. In addition, there may CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 45

have been a minor effect of Drummer including alcohol readings of greater than.009, rather than only those greater than.05 but the low values constituted only 3% of those in the study and so are not likely to have had a marked effect. These factors may have also contributed to the higher percentage of drivers with no drugs or alcohol in the current study compared to Drummer s. Table 9.1. Comparisons with other recent Australian data for drugs and alcohol in drivers in single vehicle crashes. South Australian drivers injured (n=364) 1 Percent of drivers Drivers killed 1990- Current study - 1993 (n=515) 2 drivers only (n=98) 3 No drugs or alcohol 53 37 46 Alcohol only (>0.05%) 21 40* 14 Cannabis only 7 3 4 Alcohol and cannabis 11 9 18 Stimulants 4 2 1 Benzodiazepines 3 1 3 Opiates 0 1 1 Alcohol and stimulants 1 1 1 Alcohol and benzodiazepines 1 2 0 Other drugs 6 12 1 from Longo, Hunter and Lokan (1996) 2 from Drummer (1994) Table 11 3 where toxicology was known The percentages of drivers with alcohol and cannabis were somewhat similar for the three studies. However, the current study showed a larger percentage of drivers with the combination of alcohol and cannabis. The reasons for this are unclear. 9.2 COMPARISONS WITH OTHER DATA SOURCES Based on data from earlier years, approximately 100 crashes were expected to occur in the study area during the study period. However, Victoria experienced an increase in fatal single vehicle crashes in 1996 and the total number of crashes in the study was 127. In 1990-94, 47% of serious casualty single vehicle crashes occurred in the Melbourne metropolitan area. In 1994-95, 38% of fatal single vehicle crashes occurred in the metropolitan area. In the current study, 60% of crashes occurred in the Melbourne 46 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

metropolitan area. Part of the reason for this discrepancy may be the increase in the size of the metropolitan area as a result of restructuring of municipal boundaries, and the fact that the study area did not include the whole state. The sample of crashes in the current study are compared with serious injury and fatal single vehicle crashes which occurred in Victoria in 1995 (VicRoads, 1996) in Table 9.2. In comparing the data, it should be noted that the definitions of single vehicle crashes differ slightly and that the cases in the current study do not include those occurring beyond a 200 km radius of Melbourne. Thus metropolitan crashes are somewhat over-represented in the current sample and rural crashes somewhat underrepresented. Table 9.2. Percentages of crashes by nature of crash for serious injury and fatal single vehicle crashes in Victoria in 1995 (not including struck pedestrian crashes) and the fatal single vehicle crashes in the current study (December 1995 to November 1996). Nature of crash Victoria 1995 Current study - Serious injury crashes Fatal crashes (n=128) Fatal crashes (n=127) (n=1568) Struck animal (not ridden) 2 4 0* Pole 14 13 21 Tree 30 45 49 Fence or wall 7 2 3 Embankment 6 3 2 Guide post 1 0 0 Traffic sign 1 1 0 Guard rail 3 2 2 Building 1 0 0 Fixed object 1 1 0 Traffic signal 1 2 0 Bridge (off path) 1 7 2 Traffic island 1 0 0 Roadworks 0 2 0 Safety zone 1 0 0 Struck non fixed object 3 2 3 Vehicle overturned 13 11 6 Fall in/from moving 14 4 0* vehicle Other accident 1 0 12 Total 101 99 100 * categories excluded from the current study Crashes involving trees were the most common kind of single vehicle crash in all three data sets. The severity of these crashes is demonstrated by the larger representation of crashes involving trees in both fatal data sets (45% in Victoria 1995, 49% in the current study) compared to serious injury crashes (30%). There were no fall in/from moving vehicle crashes or struck animal (not ridden) in the current study because they were not within the criteria for notification for the study. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 47

48 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

10. POSSIBLE IMPROVEMENTS IN INVESTIGATION PROCEDURES The third aim of this study was to identify improvements in procedures for the investigation of road deaths and life threatening injuries. This study arose from concern by the Victoria Police and the State Coroner that there was a need for a better understanding of the causative factors in single vehicle run-off-road crashes. Traditionally, Police investigation of these crashes has often been rudimentary when the driver has been killed (and so prosecution is not possible). Since driver-killed crashes comprise about three-quarters of all fatal single-vehicle crashes, the lack of detailed investigation of these crashes has prevented the development of a clear understanding of the factors involved. Improvements are suggested in procedures relating to alcohol, other drugs, recording vehicle information and estimation of driver fatigue. Implementation of some of these improvements will require not only changes to procedures but legislative change or improved resourcing. Such changes may be possible in the context of the proposed review of the Road Safety Act, the Parliamentary Road Safety Committee s Report into Drugs (Other Than Alcohol) In Driving and the review of the State accident data system. Despite potential difficulties, it is felt that the improvements are needed to make data collection procedures more effective from both the legal and injury prevention frameworks. 10.1 ALCOHOL Despite advances made in recent years, alcohol remains the most significant driver impairment factor in road crashes. Yet, availability of alcohol data for crashes in general suffers from such a degree of missing data that most analyses are conducted using a surrogate measure (the so-called high and low alcohol hours ) rather than recorded BAC values. Availability of BAC information for the 24 surviving drivers in the current study was complicated somewhat by ongoing criminal investigations. While the information was supplied for the study upon request, it is unclear whether this information would be available for the State Traffic Accident Record in the short term or the long term. We would propose that this issue be investigated and, if necessary, a procedure be developed to incorporate the BAC data into the State Traffic Accident Record at an appropriate time (perhaps after completion of criminal procedures). Even in the current study, BAC information was missing for 5% of drivers or riders. Each of these drivers was deceased. The State Coroner s Office stated that the information was unavailable because the family had objected to autopsy. While the wishes of the families are understandable, the prevention of further, similar deaths would be assisted by some changes to the current procedures. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 49

It is suggested that the family s right to object to autopsy be maintained but that the authorities have the right to take and analyse a blood sample from the deceased. 10.2 OTHER DRUGS The issues relating to drugs other than alcohol in crashes continue to be of concern. One of the major issues is the lack of information about the drugs consumed by drivers in crashes. Current legislation requires that blood samples be taken for testing for alcohol when injured drivers and riders present at hospital. These blood samples may also be tested for other drugs. None of the blood samples from surviving drivers in the current study appeared to be tested for other drugs and it is assumed that testing for other drugs is infrequent. Given the lack of information on drugs in nonfatal crashes, perhaps testing of all (or a given proportion) of blood samples for drugs should be undertaken for a specified period to increase our knowledge in this area. This recommendation is in line with the recommendations of the Parliamentary Road Safety Committee Inquiry into Drugs (Other than Alcohol) in Driving. While it was not a major issue in this study, the legislation does not require that drivers who are uninjured in fatal crashes have a blood sample taken. In some particular types of fatal crashes (e.g. those involving trucks and those involving pedestrians), the driver is often uninjured and so blood samples are generally not taken. While breath alcohol may be measured, this does not supply any information about the role of other drugs in these crashes. It is recommended that blood samples of all drivers in fatal crashes (whether or not injured) be taken and analysed for both alcohol and other drugs. An alternative would be to have blood samples taken and analysed whenever the breath shows alcohol (building on the strong relationship between alcohol and cannabis). While cannabis was detected in 17% of drivers and riders in the current study, this was most commonly by means of the metabolite, carboxy-thc. As mentioned earlier in the report, this does not show that the driver or rider was influenced by cannabis at the time of the crash, only that they had smoked it some time in the past several weeks. It is recommended that testing for the active ingredient, rather than the metabolite, of cannabis be undertaken, at least for fatal crashes, in the short term. Improvements in technology may reduce costs etc and allow this to be extended to nonfatal crashes in the future. 50 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

10.3 RECORDING VEHICLE DATA Monitoring of the primary and secondary safety performance of vehicles would be enhanced if sufficient details of crashed vehicles were recorded to infer if particular features are present (e.g. airbags). It is recommended that Vehicle Identification Numbers (VINs) be recorded to allow better identification of makes and models of vehicles and judgements of the presence of safety features. Currently, VINs could be added to the State Traffic Accident Record routinely by interrogating the registration data, using the vehicle registration number. In the future, automatic capture of VINs at the crash site may replace this method. In addition, for the next couple of years, it would be desirable if deployment of airbags could be recorded. Difficulty exists in determining the transitory characteristics of the vehicle, the most important of which is the speed profile prior to the crash. Others which may be less important to crash causation but should be easier to assess are related to ventilation and heating, use of cruise control etc which can be inferred from examination of the vehicle (settings of knobs etc) after the crash. Unfortunately this information was not recorded for many crashes in the current study. It is unclear whether this resulted from difficulty in ascertaining the information or a perceived lack of importance of the information. It is probably preferable for such items to be collected in special studies, rather than to burden the Police with additional workload for every reported crash. 10.4 AIS ATTENDANCE CRITERIA The AIS attended and investigated every fatal single vehicle crash within the study area for the purposes of this study. Normally, they would have only investigated those crashes which fell within their attendance criteria. These criteria include crashes in which: culpability may be involved three or more people are killed a Police officer is involved a truck or bus is involved While resource limitations are an important issue which needs to be considered, it may be useful from the point of view of training staff and developing a high quality knowledge base for prevention, to have a special focus on particular crash types (or perhaps even locations) for a particular period. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES 51

REFERENCES Drummer, O.H. (1994). Drugs in drivers killed in Australian road traffic accidents. The use of responsibility analysis to investigate the contribution of drugs to fatal accidents (Report No. 0594). Melbourne: Victorian Institute of Forensic Pathology, Department of Forensic Medicine, Monash University. Drummer, O.H. (1995). A review of the contribution of drugs in drivers to road accidents. In The Inquiry into the Effects of Drugs (Other Than Alcohol) on Road Safety, First Report. Parliament of Victoria, Road Safety Committee. Longo, M.C., Hunter, C.E. and Lokan, R.J. (1996). The incidence and role of alcohol, cannabinoids, amphetamines and benzodiazepines in non-fatal crashes. In Road Safety Research and Enforcement Conference 1996. Conference Proceedings. pp. 135-141. Vic Roads (1996). Road traffic accidents involving serious casualties - Victoria 1995. 52 MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

APPENDIX 1: CRASH QUESTIONNAIRE CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES

APPENDIX 1: QUESTIONNAIRES ESTIMATION OF RISK FACTORS FOR FATAL SINGLE VEHICLE CRASHES

MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

ESTIMATION OF RISK FACTORS FOR FATAL SINGLE VEHICLE CRASHES

MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

ESTIMATION OF RISK FACTORS FOR FATAL SINGLE VEHICLE CRASHES

MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

ESTIMATION OF RISK FACTORS FOR FATAL SINGLE VEHICLE CRASHES

MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

ESTIMATION OF RISK FACTORS FOR FATAL SINGLE VEHICLE CRASHES

MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

ESTIMATION OF RISK FACTORS FOR FATAL SINGLE VEHICLE CRASHES

MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

ESTIMATION OF RISK FACTORS FOR FATAL SINGLE VEHICLE CRASHES

MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

ESTIMATION OF RISK FACTORS FOR FATAL SINGLE VEHICLE CRASHES

MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

ESTIMATION OF RISK FACTORS FOR FATAL SINGLE VEHICLE CRASHES

APPENDIX 2: CODING MODIFICATIONS TO VEHICLES Following discussion with Vic Roads Vehicle Safety Standards section, a set of categories were developed for the question, Have any modifications been made to the vehicle?. The categories for coding modifications are: 1. common modifications These include adding audio equipment, mudflaps, towbar, cruise control, conversion to gas, etc. 2. suspension or steering modifications These include changes to the wheels or tyres (e.g. fitting alloy or chrome wheels), fitting different shock absorbers, lowering the vehicle. 3. engine or driveline modifications These include changes to the gearbox, converting an automatic to a manual or vice versa, changes to the differential. 4. structural modifications These include changing the vehicle to a convertible or a stretched vehicle, cutting into the body to make wheel tubs, adding additional seats to the vehicle. It was stated that many of these modifications may have little effect on the likelihood of the vehicle crashing or the level of injury. What they may reflect are the characteristics of the driver. CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES

APPENDIX 3: SUMMARY OF CASES. SMV Date Time Location Road type Crash Vehicle Year Deceased 1 1.12.95 0511 Oakleigh Primary arterial SEC pole Car? Driver 2 7.12.95 2205 Monbulk Primary arterial Parked car/rollover Car 1986 Driver 3 16.12.95 2340 Maribyrnong Primary arterial Fence Car 1981 Driver 4 17.12.95 0220 Near Little River Freeway Tree/fire Car 1989 Driver 5 Case w/d 6 26.12.95 1320 Patterson Lakes Freeway Bridge pylon Car 1979 Driver 7 25.12.95/ 2200- Dromana Primary arterial Tree Car 1983 Driver 26.12.95 0555 8 22.12.95 0930 Dunolly Secondary/local (?) Trees Car 1989 Driver 9 31.12.95 0300 Pascoe Vale Sth Freeway Railing/embank Car 1990 Driver 10 30.12.95 1330 Sandringham Secondary arterial Tree Car 1980 Driver 11 2.1.96 0500 Pomborneit Primary arterial Fence/rollover Car 1992 Driver 12 Case w/d 13 Case w/d 14 26.12.95 0615 Buangor Primary arterial Trees Car 1980 Driver 15 8.1.96 0735 Seaford Primary arterial SEC pole Truck 1987 Passenger 16 8.1.96 0735 Panton Hill Primary arterial Tree Car 1995 Driver 17 10.1.96 0630 Mount Martha Freeway Tree Car 1988 Driver 18 10.1.96 0905 Baxter Primary arterial Tree M/cycle 1986 Rider 19 12.1.96 1230 Haddon Secondary/local (?) Tree Car 1982 Driver 20 14.1.96 0130 Moorooduc Sth Primary arterial Tree Car 1995 Driver 21 17.1.96 1555 Rowville Primary arterial Tree Car 1985 Passenger 22 20.1.96 1809 Clyde Secondary arterial Utility pole Car 1970 Driver 23 20.1.96 1630 Nalinga Primary arterial Tree Car 1995 Passenger 24 27.1.96 0705 Werribee South Collector road SEC pole Car 1973 Driver 25 Case w/d 26 18.1.96 0440 Avenel Freeway Tree Car 1995 Passenger 27 2.2.96 0550 Barrabool Local road Rollover Truck 1995 Driver 28 2.2.96 1315 Violet Town Freeway Tree Car 1995 Driver 29 4.2.96 1750 Hunter Local road Off-road Car 1976 Driver and 2 pass. 30 8.2.96 1750 Near Foster Primary arterial Off-road Car 1989 Driver 31 8.2.96 1510 Highton Secondary arterial Off-road Car 1994 Driver 32 10.2.96 1650 Bulleen Int. prim. & second. Utility pole Car 1971 Driver 33 8.2.96 1342 Moorium Primary arterial Tree Truck 1995 Driver 34 10.2.96 0430 Castella Primary arterial Tree Car 1987 Driver 35 11.2.96 1030 Beaufort Secondary arterial Tree Car 1983 Driver 36 11.2.96 2230 St Kilda Secondary arterial Kerb then car M/cycle 1986 Rider 37 15.2.96 0300 Hawthorn Primary arterial SEC pole Car 1975 Passenger 38 16.2.96 2050 Berwick Collector road Off-road Car 1982 Passenger 39 19.2.96 0925 Taggerty Local road Rollover Car 1984 Passenger 40 Case w/d 41 25.2.96 1750 Coldstream Primary arterial Tree Car 1976 Passenger 42 25.2.96 2355 Mt Bute Secondary arterial Rollover Car 1995 Passenger 43 28.2.96 0125 Noble Park Primary arterial Tree Car 1974 Passenger 44 28.2.96 2050 Noble Park Primary arterial Tree Car 1988 Driver 45 3.3.96 0315 Upwey Primary arterial SEC pole Car 1973 Driver 46 10.3.96 0033 Epping Primary arterial Tree Car 1979 Driver 47 16.3.96 1525 Warragul Freeway Tree Car 1988 Passenger 48 16.3.96 2330 Footscray Primary arterial Traffic signal pole Car 1986 Passenger MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

SMV Date Time Location Road type Crash Vehicle Year Deceased 49 22.3.96 1420 Wonga Park Local road Rollover Car 1988 Passenger (unsealed) 50 29.3.96 1925 Lilydale Primary arterial Tree Car 1974 Driver 51 1.4.96 1110 Moe Freeway Off road Car 1989 Driver 52 5.4.96 2320 Bulleen Primary arterial SEC pole Car 1978 Driver and passenger 53 6.4.96 0308 Prahran Primary arterial Tree Car 1981 Driver 54 9.4.96 1200 Hurstbridge Primary arterial Ped light pole Car 1978 Driver 55 12.4.96 0150 Bendigo Local road Tree Car 1983 Passenger Passenger 56 14.4.96 1240 Longwarry North Local road Off road Car 1976 Driver 57 17.4.96 1100 Lang Lang Minor arterial Tree Car 1994 Driver 58 20.4.96 0255 Nagambie Primary arterial Tree Car 1977 Driver 59 21.4.96 0415 Richmond Freeway Off road M/cycle 1989 Pillion 60 5.4.96 2115 Mitcham Primary arterial Tree Car 1977 Driver 61 23.4.96 0018 Mount Duneed Primary arterial SEC pole Car 1956 Driver 62 25.4.96 2228 Springvale Primary arterial Tree Car 1987 Driver 63 27.4.96 0515 Box Hill South Primary arterial Tree Car 1993 Driver 64 27.4.96 0930 Tatura Minor arterial/local SEC pole Car 1989 Driver 65 4.5.96 0210 Blairgowrie Primary arterial Tree Car 1986 Driver 66 4.5.96 2353 Kinglake Local road Tree Car 1994 Driver 67 5.5.96 1500 East Doncaster Local Tree Car 1984 Driver 68 6.5.96 1400 Collingwood Local Brick fence Car 1982 Driver 69 12.5.96 0354 Glen Waverley Freeway Pylon/rollover Car 1992 Driver 70 15.5.96 1825 Preston Primary arterial Tree Car 1985 Driver 71 16.5.96 0810 Toolern Vale Minor arterial Off road Car 1993 Driver 72 13.2.96 2250 Clayton Local SEC pole Car 1979 Passenger 73 15.5.96 1150 Cowes Local Tree Car 1984 Driver 74 3.6.96 0930 Ringwood North Primary arterial Probable heart att. Car 1989 Driver 75 3.6.96 1645 Whittlesea Primary arterial Fence Car 1978 Driver 76 9.6.96 0500 Corio Freeway Tree Car 1980 Driver 77 10.6.96 1050 Euroa Local road Tree Car 1972 Driver 78 16.6.96 0348 Carlton Collector road Parking meter Car 1990 Passenger 79 22.6.96 1310 East Kew Secondary arterial Sewerage vent Car 1991 Driver pipe 80 22.6.96 2045 Kerrisdale Primary arterial Tree Car 1976 Driver 81 22.6.96 2115 Langwarrin Secondary arterial Rollover Car 1981 Driver 82 27.06.96 0700 Norlane Primary arterial Pole M/cycle 1986 Rider 83 06.07.96 0245 Ashburton Primary arterial Tree Car 1981 Passenger 84 15.07.96 0310 Windsor Primary arterial Power pole Car 1995 Driver 85 15.07.96 0745 Loch Local road Embankment Car 1974 Driver 86 16.07.96 1426 Sedgwick Local road Tree Car 1986 Driver 87 19.07.96 0440 Tynong Freeway Tree Car 1974 Driver 88 22.07.96 1656 Officer Freeway Tree Car 1986 Driver 89 26.07.96 1310 Avenel Freeway Traffic sign Car 1993 Driver 90 06.08.96 0142 Caulfield Secondary/collector Parked car, pole, Car 1986 Driver sign 91 05.08.96 2302 West Melbourne Secondary/secondar Tree Car 1990 Passenger y 92 07.08.96 1615 Silvan South Secondary arterial Tree Car 1982 Passenger 93 29.07.96 0955 Kerrisdale Primary arterial Tree Car 1990 Driver CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES

SMV Date Time Location Road type Crash Vehicle Year Deceased 94 07.08.96 1515* Undera North Local unsealed Tree M/cycle 1989 Rider 95 08.08.96 2225 Strathmore Collector Pole Car 1989 Driver and passenger 96 10.08.96 1005 Forrest Minor arterial Tree Car 1971 Driver 97 12.08.96 1815 Creswick Primary arterial/local Rollover Truck 1967 Passenger 98 14.08.96 0600 Altona Freeway Tree Car 1987 Driver 99 25.08.96 0130 Templestowe Primary arterial SEC pole Car 1978 Passenger 100 25.08.96 0300 Wendouree Local Tree Car 1995 Driver 101 Case w/d 102 31.08.96 0455 Mt Waverley Primary arterial Pole Car 1994 Driver 103 01.09.96 1304 Macleod Local road Pole M/cycle 1990 Rider 104 05.09.96 2210 Shepparton Local road Tree M/cycle 1981 Rider 105 07.09.96 2300 Ararat Minor arterial Tree Car 1993 Driver 106 08.09.96 0415 Mandurang South Local road Tree Car 1979 Passenger 107 12.09.96 2315 Warragul Local road Tree Car 1975 Passenger 108 13.09.96 2110 Deer Park Primary arterial Tree Car 1985 Driver 109 13.09.96 1215 Thornton Local road SEC pole Truck 1971 Driver 110 19.09.96 0125 Hawthorn Secondary arterial Tram safety barrier Car 1977 Driver 111 13.09.96 0100 Hallam Primary arterial Stationary truck Car 1998 Driver 112 21.09.96 0345 Buninyong Primary arterial Rollover Car 1985 Driver 113 22.09.96 0925 Noble Park Secondary arterial Power pole Car 1981 Driver 114 25.09.96 1355 Echuca Primary arterial Pole Car 1981 Driver 115 25.09.96 1754 Springvale Primary arterial Tree Car 1988 Driver 116 28.09.96 0300 Oakleigh Primary arterial Power pole Car 1983 Driver 117 29.09.96 2028 Box Hill North Secondary arterial Tree/fence Car 1983 Driver 118 15.10.96 1530 Brunswick Local road Parked car M/cycle 1996 Rider and pillion 119 19.10.96 0235 Balwyn Secondary/local SEC pole Car 1991 Driver 120 19.10.96 1700 Newham Minor arterial Tree Car 1977 Driver 121 28.10.96 1030 Nangana Primary arterial Culvert M/cycle 1985 Rider 122 02.11.96 2010 Keysborough Primary arterial Rollover Car 1982 Driver 123 04.11.96 1935 Laverton Off-road Drain M/cycle 1986 Rider 124 08.11.96 1550 Tallarook Freeway Tree Car 1988 Driver 125 19.10.96 2320 Croydon Secondary/local Power pole Car 1985 Driver 126 16.11.96 0240 Noorat Minor arterial Tree Car 1977 Driver 127 17.11.96 0255 Balwyn North Secondary arterial SEC pole Car 1990 Driver 128 19.04.96 1640 Tottenham Primary arterial Large rock Car 1993 Driver 129 21.11.96 0250 The Gurdies Primary arterial Embankment Car 1980 Driver 130 21.11.96 1608 Woodend Local unsealed Tree Car 1975 Driver 131 23.11.96 0005 Bullengarook Local road Tree Car 1992 Driver 132 20.11.96 2000 Craigieburn Collector/local Guard railing Car 1983 Driver 133 18.09.96 1748 Ferny Creek Primary arterial/local Tree Car 1973 Driver MONASH UNIVERSITY ACCIDENT RESEARCH CENTRE

APPENDIX 4: CONCENTRATIONS OF CANNABIS METABOLITE The data below are summarised from toxicology reports on killed drivers supplied by the Victorian Institute of Forensic Medicine. The metabolite measured in each instance was 11-Nor-9-Carboxy-Tetrahydrocannabinol. Concentration Specimen BAC Other drugs (ng/ml) 124 urine.090 57 blood.000 detected EMIT not detected SIM urine.150 opiates benzodiazepines - no level 321 urine.180 benzodiazepines 27 blood.000 40 blood.090 294 plasma.180 74 blood.250 18 blood.070 45 blood.190 42 urine.220 19 blood.240 18 urine.190 164 urine.000 700 urine.000 urine.220 detected EMIT not detected SIM - no level 63 urine.190 19 urine.220 5 urine.270 50 blood.200 110 urine 520 blood.180 54 urine 43 blood.180 CHARACTERISTICS OF FATAL SINGLE VEHICLE CRASHES