Introduction

The decision to use road vehicle collision as a suicide method could be influenced by factors such as perception of pain, likelihood of fatal outcome, ready access to vehicles, likelihood of bystander intervention, and perception that the event will be ruled as unintentional (in order to secure life insurance policy payment for family and protect family against social stigmatisation; [1]. Suicide events involving road vehicle collision involve a greater risk of serious injury and death to bystanders than other suicide methods [2, 3]. For suicide events involving vehicles with a large weight disparity (e.g., car/van and truck), almost 30% result in injury of another person [4] and almost 4% result in death of another person [5].

Road vehicle collision as a suicide method involves either a driver intentionally colliding into an object (e.g., road infrastructure, a natural fixed object, another vehicle) or a pedestrian (i.e., person on foot) purposefully placing themselves in the collision path of an oncoming vehicle. Drivers who died by suicide are relatively more common [5–8], and research has largely focused on single-vehicle collisions [9]. In Australia, only two studies have examined suicide events involving road vehicle collision, using motor vehicle collision data in Queensland. One study estimated that single-vehicle collisions constituted 55% of driver deaths by suicide between 2011 and 2015 [10]. The finding is consistent with the other Australian study that estimated 58.6% of driver deaths by suicide between 1990 and 2007 involved colliding into a tree or power pole [11]. Estimates for multiple-vehicle collisions differed in the two studies, with reports that 27.5% [11] and 80% [10] of driver deaths by suicide involved a collision with an oncoming truck.

Pedestrians who died by suicide have done so by jumping off road infrastructure or, more commonly, by stepping into the path of an oncoming vehicle [12]. The topic of pedestrian deaths by suicide has received much less focus in research than driver deaths by suicide. The limited research that has been done on the topic suggests that pedestrian deaths by suicide typically involve colliding into an oncoming heavy vehicle such as a truck [6].

Drivers who died by suicide are more often men, young to middle aged, unemployed, and not married [10, 11, 13–15]. The majority of drivers who died by suicide in Australia also consumed alcohol prior to the collision [11], and experienced legal and financial problems more so than people who used other suicide methods. These risk factors for vehicle-collision suicide mirror those reported in other countries [1, 3, 5, 7, 14–18], and are also commonly reported for other suicide methods.

There is still much that is not understood about vehicle-collision suicide in Australia. In particular, studies have yet to examine the suicide rate involving road vehicle collision over time, the proportion of different types of vehicle collision, and individual characteristics of those who died by vehicle-collision suicide. The current study addresses these gaps by analysing detailed case information from the National Coronial Information System. First, we identified all deaths by suicide involving road vehicle collision that occurred over a 17 year period (2001−2017) in Australia, during which an increase in road vehicle use was reported [19–21], and compared the annual trend to that of other suicide methods. Second, we examined the proportions of different travel modes used (e.g., car/van, on foot) and their collision counterparts (i.e., the corresponding object that the deceased collided with, e.g., tree, truck). Third, we identified individual characteristics associated with the likelihood of dying by vehicle-collision suicide and compared them to those of other suicide methods.

Materials and methods

Results

Trends in deaths by suicide involving road vehicle collision and other methods

A total of 640 individuals died by suicide in road vehicle collisions between 2001 and 2017. The travel mode used was obtained for 594 individuals and the collision counterpart was obtained for 573 individuals. Drivers accounted for 61% of people who died by vehicle-collision suicide and pedestrians who used an oncoming road vehicle for suicide accounted for the remaining 39%. Fig 1 shows an overall upward trend in the national suicide rate involving road vehicle collision over time (per 100,000 population). There was variability in the trends at the state/territory-level, with data from two states showing an increase in deaths by suicide involving road vehicle collision (Victoria and New South Wales) and the rest showing no change or a declining trend. A total of 41,890 individuals died by suicide involving other methods in the same time period. The series is characterised by an overall decline in the national suicide rate (Fig 2). State/territory-level variability in the trends was also observed, with three states showing a decline in deaths by suicide involving other methods (Victoria, New South Wales, and South Australia) and the remainder showing no change or an increasing trend.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Suicide rates involving road vehicle collision between 2001 and 2017 in Australia.

https://doi.org/10.1371/journal.pone.0299590.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Suicide rates involving other methods between 2001 and 2017 in Australia.

https://doi.org/10.1371/journal.pone.0299590.g002

Table 1 provides additional information about the increase in deaths by vehicle-collision suicide throughout the study period. It shows that the years that accounted for the increase were 2012 (IRR = 1.69), 2015 (IRR = 1.72), and 2017 (IRR = 2.05). In each of these years, the IRR was significantly higher than in 2001 (p≤0.001). In contrast, there was clear evidence of fewer deaths by suicide involving other methods in all years relative to the 2001 reference year.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Incidence rate ratios for deaths by suicide involving road vehicle collision and other methods.

https://doi.org/10.1371/journal.pone.0299590.t001

Types of road vehicle collisions

Of the drivers who died by suicide (N = 365), 264 (72%) were in single-vehicle collisions. For the remaining drivers who died by suicide (N = 101), 99 were in multiple-vehicle collisions of which 81 collided with an oncoming truck. Table 2 shows that when a car/van was driven as the travel mode for suicide, 74% of the collisions involved a fixed object (commonly a tree or concrete barrier) and 21% involved an oncoming truck. When a motorcycle was driven as the travel mode for suicide, 40% of the collisions involved a fixed object, 47% involved an oncoming truck, and 13% involved an oncoming car/van. Of the pedestrians who used an oncoming vehicle for suicide, 58% collided with a truck and 23% collided with a car/van.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Proportion of travel modes used and collision counterparts for different types of road vehicle collision as a suicide method between 2001 and 2017 in Australia.

https://doi.org/10.1371/journal.pone.0299590.t002

Characteristics of individuals who died by suicide involving road vehicle collision and other methods

Table 3 shows that the individual characteristics associated with the likelihood of dying by vehicle-collision suicide were younger age, non-Indigenous origin, and country of birth. Individuals who were younger than 25 years old had more than five times the odds of dying by vehicle-collision suicide compared with those who were 65 years and older (OR 5.27), those aged 25–34 years and 35–44 years had more than triple the odds (OR 3.76 and 3.50, respectively), and those aged 45–54 years had more than double the odds (OR 2.72). Non-Indigenous individuals had triple the odds of dying by vehicle-collision suicide compared with Indigenous individuals (OR 3.36). Individuals born overseas had greater odds of dying by vehicle-collision suicide compared with individuals born in Australia (OR 1.40).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Descriptive, univariate, and multivariate results for individual characteristics of those who died by suicide involving road vehicle collision and other methods between 2001 and 2017 in Australia.

https://doi.org/10.1371/journal.pone.0299590.t003

Discussion

The current study examined the annual suicide rate involving road vehicle collision and other methods for the period of 2001–2017, the proportion of different types of vehicle collisions, and individual characteristics associated with the likelihood of dying by vehicle-collision suicide compared with other suicide methods. We observed an overall increase in the national suicide rate involving road vehicle collision over the study period, with the rate being significantly greater in 2012, 2015, and 2017 than in 2001. The rate consistently increased between 2005 and 2008 (an aggregate of the upward trend observed in Queensland, Victoria, Western Australia, and Tasmania in the same period). In contrast, we observed an overall decline in the national suicide rate involving other methods throughout the same time period, with the rate being significantly lower in each year since 2003 compared with 2001.

It is worth considering the factors that may explain why deaths by suicide involving road vehicle collision have increased in the face of a decline in deaths by suicide involving other methods. Some substitution may be occurring, with an increasing availability of cars (road vehicle use consistency increased between 2001 and 2017 in Australia [19–21]). In addition, suicide prevention efforts that may have reduced deaths by suicide involving other methods (e.g., the annual incidence of suicide deaths by motor vehicle exhaust gas has decreased by 57% between 2001 and 2006 [23]) may have inadvertently led to an increase in deaths by vehicle-collision suicide. However, these explanations are at best only likely to partially explain the diverging trends observed in the current study. Deaths by vehicle-collision suicide continue to account for a small proportion of all deaths by suicide, these relatively small numbers make it difficult to definitely determine the cause of their changing patterns.

We found that most suicide events involving road vehicle collision were by drivers, which is consistent with studies conducted in Finland [5], Scotland [6], Sweden [8], and Switzerland [7]. Of the drivers who died by suicide, 72% were in single-vehicle collisions, which is greater than 55% [10] and 59% [11] reported in earlier Australian studies. Car/van was the predominant mode of travel used and motorcycles were occasionally used, with the most common collision counterpart being a tree. For the remaining drivers who died by suicide, 27% were in multiple-vehicle collisions. Of these drivers, 82% collided with an oncoming truck. This estimate is consistent with an earlier Australian study [10] and considerably greater than the 28% reported in another Australian study [11]. However, these earlier studies used comparatively smaller sample sizes than the current study (N = 10 and 52 vs. N = 365). In addition, we found that more than one-third of suicide events involving road vehicle collision were by pedestrians, with the most common collision counterpart being an oncoming truck. This estimate is consistent with an earlier study in Scotland which also used a smaller sample size (N = 17).

We identified individual characteristics associated with likelihood of dying by vehicle-collision suicide in comparison to other suicide methods. Individuals younger than 25 years old had the highest likelihood of dying by vehicle-collision suicide among all age groups. Specifically, they were five times more likely to die by vehicle-collision suicide than individuals aged 65 years and older. Those aged 25–44 years and 45–54 years were more than three times and two times more likely to die by vehicle-collision suicide, respectively, than individuals aged 65 years and older. Males were more likely to die by vehicle-collision suicide than females. This supports earlier Australian studies that reported drivers who died by vehicle-collision suicide are often by younger men [10, 11, 13, 14], which is a finding that is also observed in other countries [1, 3, 5, 7, 14–18]. Our findings also indicated that individuals who were non-Indigenous were three times more likely to die by vehicle-collision suicide than those who were Indigenous, and individuals born overseas were more likely to die by vehicle-collision suicide than those born in Australia. Unemployed individuals were not more likely to die by vehicle-collision suicide than those employed and not in the labor force, in contrast to an earlier Australian study [11].

Different types of initiatives are available to reduce suicide events involving road vehicle collision. First, general road safety measures for preventing unintentional road deaths are likely to have suicide preventive benefits. For example, setting reduced speed limits may reduce pedestrian deaths by vehicle-collision suicide (pedestrian fatality risk rapidly increases with any small increase in vehicle-pedestrian impact speed between 30–70km/hr [24]). Installing cable barriers and guardrails on both sides of a road and between opposite direction lanes may also reduce driver deaths by suicide in single-vehicle collisions and multiple-vehicle collisions, respectively (cf. concrete barriers and ditches which are associated with a relatively higher risk of injury in road vehicle collisions [25]). However, Australia has a vast road network and road vehicles are widely accessible. Initiatives that involve the installation of infrastructure (e.g., signs, road safety barriers, crisis telephones, and CCTV) may face feasibility and cost effectiveness issues, and would likely be challenging to implement especially over a large area. Technological innovations in vehicles that improve general road safety may also prevent suicide events involving road vehicle collision. An earlier Australian study found that most drivers who died by vehicle-collision suicide had consumed alcohol prior [11]. Furthermore, the relationship between alcohol use and increased predisposition toward suicidal behavior is well known [26–28]. Alcohol interlocks in vehicles can have a role in restricting vehicle access to an at-risk individual who is under the influence of alcohol and intend to self-harm. ‘Smart’ interlocks can also be adopted to prevent an at-risk driver from dying by suicide via remote lockdown of a vehicle to secure the individual inside long enough for assistance to arrive. In addition, vehicle systems can monitor driving behaviors that correspond to those of drivers who died by suicide in vehicle collisions (e.g., abrupt sharp steering, excessive speeding, not wearing a seat belt) to trigger for reduction of speed and warning signals for oncoming vehicles.

Second, road vehicle collision as a suicide method should be the focus of targeted interventions. For example, refining media guidelines to specifically advise on responsible reporting of road vehicle collision fatalities [29] may reduce the risk of so-called “copycat” behavior [30, 31]. For social media (e.g., Facebook, Twitter), amending platform guidelines and releasing platform updates can help to reduce the spread of harmful content, promote help-seeking behavior, and direct at-risk users to crisis support services (e.g., Lifeline). Based on our findings, such efforts might preferentially target some groups (e.g., males younger than 25 years old, individuals who were non-Indigenous and born overseas). Our findings also indicate that oncoming trucks were involved in most pedestrian deaths by suicide and driver deaths by suicide in multiple-vehicle collisions. Therefore, enhancing national assessment practices for heavy vehicle licensing (e.g., vehicle-collision suicide awareness training, defensive driving training, simulations of suicidal behavior in hazard perception testing) may reduce the risk of death from vehicle collision suicide by pedestrians and drivers. The installation of signs to display positive messaging and information about help services, crisis telephone support services to increase opportunities for help-seeking, and blue light-emitting diode lamps (which has been shown to reduce suicide deaths at railway stations with no substitution at neighbouring stations [32]) may also reduce suicides by vehicle collision, especially at areas where such suicide events are frequent. Last, general primary prevention of suicide, especially measures targeting adolescents (i.e., school-based programs), are likely to also reduce suicides by road vehicle collision. These involve problem-solving training, basic mental health education, encouragement of help-seeking behavior, gatekeeper training, and public awareness campaigns that empower individuals to talk about suicide [33, 34].

Our study examined the most representative dataset on suicide deaths involving road vehicle collision in Australia to date, compiling coronial data from each Australia state and territory across a 17-year period (2001–2017). To our knowledge, this is the most comprehensive study on road vehicle collision as a suicide method in Australia, and the first to report its trend over time. Our study is also the first to examine pedestrians who died by suicide involving road vehicle collision in Australia.

Nevertheless, our conclusions are limited by the likely underestimation of the true number of drivers who died by suicide. Underestimation of drivers who died by suicide is widely acknowledged [5–7, 9, 14, 15, 18], and is due to a range of difficulties faced by coroners in making an explicit determination of intent. Sufficient evidence on critical background and contextual information about the deceased may at times be unavailable [2, 10, 11, 18, 35–39]. This includes indicators of psychological condition (e.g., history of suicidal behavior and ideation), emotional state (e.g., a suicide note), and driver behavior prior to the collision (e.g., eyewitness accounts). A suicide verdict requires a standard of proof with a high degree of certainty, which naturally vary among coroners even given the same key information [39–45]. The result is potential misclassification of drivers who genuinely died by suicide as accidents [9, 18, 46, 47] and due to undetermined causes (which remain as open verdicts and are not in any official statistics). Underestimation of the true number of drivers who died by suicide has been acknowledged in studies conducted in Australia [11] and other countries [5, 46, 47], and could explain the large discrepancy in vehicle-collision suicide rates among other countries in the past five decades (between 1.6% and 32% of all road fatalities [1–3, 5–8, 15, 16, 48–50]).

Future research should focus on identifying immediate precursors to suicide events involving road vehicle collision. These include blood alcohol level (most drivers who died by suicide in Queensland, Australia, were intoxicated prior to the collision, more so than other methods of suicide combined [11]), presence of illicit substances, existing mental health issues, and exposure to recent adverse life events (e.g., legal, financial, and marital problems). Subsequent studies should also focus on identifying specific areas where deaths by suicide involving road vehicle collision are frequent. This would enable the implementation of interventions at problematic sites (e.g., signs, road safety barriers, blue light-emitting diode lamps) that may reduce deaths by vehicle-collision suicide.

Conclusions

This research provides a longitudinal profile of road vehicle collision suicide in Australia. Our conclusions are: (i) the vehicle-collision suicide rate increased over time while the suicide rate by other methods decreased; (ii) vehicle-collision suicide events were predominantly by drivers in single-vehicle collisions and a substantial minority by pedestrians; (iii), individuals who die by vehicle-collision suicide are more likely to be young males, non-Indigenous, and born overseas when compared to individuals who die by other methods. Our findings inform future research directions and prevention initiatives for road vehicle collision suicide in Australia.

Acknowledgments

We would like to thank the NCIS as the database source of data, and the Victorian Department of Justice and Community Safety as the organization source of data.