Identification of key risk factors related to serious road injuries and their health impacts

Deliverable 7.4 of the H2020 project SafetyCube
Auteur(s)
Reed, S.; Weijermars, W, et al.
Jaar

Because of their high number and slower reduction compared to fatalities, serious road injuries are increasingly being adopted as an additional indicator for road safety, next to fatalities. Reducing the number of serious road injuries is one of the key priorities in the EU road safety programme 2011-2020. In 2013, the EU Member States agreed on the following definition of serious road traffic injuries: a serious road traffic injury is a road traffic casualty with a Maximum AIS level of 3 or higher (MAIS3+).

One recommendation created by the EU SUSTAIN project was to conduct “A more detailed study of the causes of serious road injuries, [which] could reveal more specific keys to reduce the number of serious injuries in the EU”. This recommendation is addressed through the identification of crashrelated causation and contributory factors for selected groups of casualties with relatively many MAIS3+ casualties compared to fatalities and groups with a relatively high burden of injury of MAIS3+ casualties.

This deliverable is made up of two parts brought together in order to determine the main contributory factors detailed above. This two-step approach initially identifies groups of casualties that are specifically relevant from a serious injury perspective using national level collision and hospital datasets from 6 countries. Following the determination of groups of interest a detailed analysis of the selected groups using indepth data was conducted. On the basis of in-depth data from 4 European countries the main contributory and causal factors are determined for the selected MAIS3+ casualty groups.

Alongside the three proceeding deliverables that have formed the major outputs of WP7, deliverable D7.4 is aimed at addressing serious injury policy at an EU levels. As such this report is broadly aimed at policy makers although the inclusion of results from in-depth data analysis also provides information relevant to stakeholders, particularly those working in vehicle design and manufacture or road user behaviour.

SELECTION OF GROUPS OF MAIS3+ CASUALTIES FOR FURTHER ANALYSIS (STEP 1)

The objective of this analysis was to select the relevant groups of serious road injuries (MAIS3+ casualties) considering number of casualties and health impacts. Relevant groups of casualties are groups with:

  • A relatively large number of MAIS3+ casualties, in relation to fatalities
  • Relatively large health impacts, quantified by Years Lived with Disability (YLD) in relation to Years of Life Lost (YLL)

The selection was based on hospital discharge register data and road fatality registers from the following countries: England, The Netherlands, Rhône region in France and Spain. In addition, some analyses were done on hospital discharge and other injury data from Austria and GIDAS data from Germany. Casualties were grouped according to transport mode, age and gender and EUROCOST injury group. The distribution of casualties over transport modes, age and gender and over EUROCOST injury groups was compared between fatalities and MAIS3+ casualties and between YLL and YLD. Moreover, MAIS3+ to fatality ratios and YLD to YLL ratios were calculated for each transport mode, each combination of age and gender and each EUROCOST injury group.

The following groups of casualties are overrepresented among MAIS3+ casualties and/or the burden of injury of these casualties compared to fatalities and are therefore selected for further in-depth analysis to determine risk factors:

  • Cyclists: In all countries, cyclists show the highest MAIS3+/fatality ratio and YLD/YLL ratio of all transport modes. Further analysis of national crash statistics data shows that cyclists are often injured in crashes without motorized vehicles and that the most common types of injury obtained by MAIS3+ casualties among cyclists are skull-brain injuries other than concussions, open head wounds and facial injuries and hip fractures.
  • 0-17 yrs; this age group shows a relatively large share in the number of MAIS3+ casualties and the burden of injury of these casualties. Moreover, the average burden per casualty is relatively high for these casualties, due to a long remaining life expectancy. Further analysis of national crash data shows that 0-17 yrs MAIS3+ casualties are relatively common among pedestrians and to lesser extent cyclists and that the most relevant injuries are skull-brain injuries other than concussions, open head wounds and facial injuries and femur shaft and knee/lower leg fractures. Another age group that could be considered relevant from a burden of MAIS3+ injury perspective are casualties of 50 yrs and older. These casualties show a somewhat higher share in YLD compared to YLL in England and in the Netherlands, but not in the other countries.
  • Spinal cord injuries; these MAIS3+ injuries always result in long-term disabilities and therefore are a main contributor to years lived with disability of MAIS3+ casualties in all four included countries. These injuries are relatively common among car occupants
  • Knee/lower leg fractures; This EUROCOST injury group has the highest share in the burden of injury of MAIS3+ casualties in the Rhône region and in Spain and is the second largest group in England. Moreover, in all countries, the share in YLD is higher than the share in fatalities for these fractures. Further analysis on the basis of national crash statistics shows that knee/lower leg fractures are most common among powered two-wheelers and are relatively common among younger road traffic casualties. Also for femur shaft fractures, the share in YLD is higher than the share in fatalities. However, due to time constraints, this group was not selected for further analysis.

As not all EU-member states currently hold data on MAIS3+ casualties, police reported serious road injuries were also investigated to determine whether they provide a good picture of the distribution of serious road injuries over transport modes and age and gender. This appears not to be the case. Due to a difference in injury severity definition and underreporting by the police, police reported serious road injuries show a different distribution over transport modes and age groups than hospital reported MAIS3+ casualties. MAIS3+ casualties among cyclists are for example heavily underreported by the police and therefore the share of cyclists is substantially underestimated when police reported injuries are used.

Finally, a comparison of MAIS3+ to fatality ratios between the countries shows that ratios differ considerably between countries. As a result, MAIS3+ to fatality ratios cannot easily be transferred from one country to another. Consequently, it does not make sense to derive MAIS3+ to fatality ratios on a European level.

ANALYSIS OF GROUPS OF MAIS3+ CASUALTIES IDENTIFIED (STEP 2)

For this analysis, in-depth collision investigation datasets were applied, as they contain more detailed information relating to the specific causation factors that lead to collisions or the mechanisms that contribute towards injuries; both of these can be considered weaknesses of national level datasets to some extent. Four of the five groups recommended by ‘step 1’ were taken forwards for analysis in this report, a decision to cover only four groups was made as a matter of expediency due to time constraints.

The guidelines for the in-depth analysis stage were kept very open; this was predominantly due to the differences in the data collection, sampling and storage of information in the various datasets, but also to provide the national in-depth dataset experts a free reign over how best to provide and present results specific to the causation and contributory factors for the groups of interest. Analysis was conducted using data from four countries; data which is representative of the national picture from Germany and England (both in-depth), and regional data from Barcelona, Spain and The Netherlands (linked hospital and police data and in-depth respectively).

Combining the in-depth datasets for the four countries involved in the in-depth analysis resulted in over 70 thousand individual collisions. In total and across all groups identified through the initial selection process just over 1% of this dataset was analysed. This indicates that groups that have relatively large numbers of MAIS3+ casualties in relation to fatalities, or relatively large health impacts in relation to Years of Life Lost exist in very small numbers within current and historic in-depth datasets.

In total data from four countries has been used resulting in an indication of the associated risk factors. The word ‘indication’ is used as it is not possible to provide statistically robust results for all the groups of interest across all countries, subsequently the results should be seen as a window into each group rather than a complete answer or solution to a problem.

The conclusions drawn from the data analysis show that:

For cyclists of all age groups the primary contributory factor as determined through the independent collision investigation is attributed to the other road user in more than half of the sample. Collisions involving a broad description of ‘entering or crossing a priority road’ are the most prevalent accident type with crashes where a cyclist does not give way, or get given right of way in a traffic situation most common. Some of these priority related crashes involve access to property rather than road junctions.

Over 90% of the causes attributed to cyclist collisions were ‘human’ factors such as distractions, emotions, attention, and physiological conditions. Factors of perception (expecting, looking and planning), legal (disobeying signs or signals, unfit to drive due to alcohol or drugs) and attention (distraction, inattention) were much more prevalent for road users whose actions initiated the collision event, particularly for injudicious actions attributed to cyclists. Besides the causation factors associated with driver or rider errors, crossing crashes are commonly associated with visibility/vision obscuration, road infrastructure factors and legal issues/infringements. Causation factors related to parked or moving vehicle or issues related to infrastructure/road geometry.

Single vehicle cycle accidents are the second most common collision type across all country data but this shows regional differences, for example, single cycle collisions represented only 2% of English sample but 28% and 50% in the German and Netherlands datasets respectively. Distraction of the cyclist was a factor in around 20% of the single vehicle accidents, leading to cyclists colliding with other road users or objects.

For 0 to 17 yrs road users, pedestrians and passengers of cars were the two largest groups in the sample of 0-17 yrs road users sustaining MAIS 3+ injuries. Cyclists and motorcycle riders were less evident in these data. The sample road users were most often involved in an accident as a pedestrian. They have a higher share of accidents when crossing the road as a pedestrian and have less driving accidents.

Across all road user types, crashes involving some form of crossing or turning were slightly over represented compared to collisions involving road users of other age groups. Riders of motorcycles and cyclists aged 0 to 17 yrs are only present in turning and crossing type collisions and crossing accidents were more frequently found among 0 to 17 yrs road users with MAIS 3+ injuries than among older road users with MAIS 3+ injuries.

Considering the causations, for drivers involved in a collision with a 0 to 17 yrs road user the most common causation factors were (i) exceeding the speed limit (ii) failed to look properly and (iii) distraction. Breaking down the 0 to 17 yrs group by road user type shows that common causal factors for cyclists were; (i) failed to judge vehicle path or speed, (ii) careless/reckless or in a hurry, and (iii) inexperience. Additional relevant causal factors related to perception (expecting, looking and planning) and conflict (interpersonal communications) were more prevalent for the 0 to 17 yrs cyclist sample than that in collisions involving other older road users.

The most common causation factors for pedestrians aged 0 to 17 yrs are related to the broad groups of ‘perception’ and ‘conflict’. Additionally a range of more specific factors, involving ‘information admission’ factors such as a wrong focus of attention or attention hindered due to physiological conditions (includes factors such as alcohol and drugs alongside medical conditions and physical stress or fatigue) were present The analysis of the accident causes reveals that most of the failures attributed to collisions involving a young road user are based on human failures, this includes a range of more specific factors, including ‘information admission’ factors such as a wrong focus of attention of attention hindered due to physiological conditions. Very few causes from the vehicle or the environment were found. This group did involve environmental causation factors which were almost completely limited to visual obstructions.

The majority of persons that suffered spinal cord injuries were car occupants in collision with fixed objects or other vehicles and within this sample, rollover crashes appear to be over-represented compared to other types of crashes. In rollover crashes all of the spinal cord injuries were from contact with an intruding roof.

Other impact types (front/rear/side) were underrepresented in the MAIS 3+ spinal cord injury sample, compared to those involved in other severe collisions. In general, for rear and side impacts, high levels of vehicle crush and damage was seen. A similar picture was also apparent in frontal collisions where high levels of crush were seen from narrow impacts with small overlaps. The analysis into the cause of the car occupant spinal cord injuries in non-rollover crashes shows that they were mostly caused by the body movement and not a result of contact or direct trauma to the neck/spine itself. In the sample of MAIS 3+, car occupants did not have a high number of injuries recorded but the type and severity of these when they occurred were some of the most severely injured occupants compared to the average injury score. All spinal cord injuries irrespective of the type of collision were located in a small region of the upper spine between C1 and T1.

Motorcycle riders and cyclists also sustained spinal cord injuries but in smaller numbers and with mixed injury causes. Not surprisingly, cyclists and PTW riders were more likely to have a spinal cord injury from an external object such as the road surface, off road surface or road side furniture.

Riders of powered two-wheelers (PTWs) had the largest share of serious injuries within the MAIS 3+ Knee and lower leg fractures sample. Severely injured car occupants and pedestrians also often suffered lower leg or knee injuries, though not as frequently as PTW riders. Although evident in the sample cyclists who received MAIS3+ knee or lower leg fractures were not as prevalent.

Vulnerable Road Users (PTWs, pedestrians, cyclists) most frequently suffered lower leg or knee injuries in a collision with a car, whereas car occupants more often sustained their injuries from a collision with a fixed object such as a tree or pole. The lower leg or knee injury was most frequently caused by impact with the front bumper of the
collision opponent amongst Vulnerable Road Users. Additionally, PTW riders commonly received these injuries from impact with the front wing, rear bumper, or had collisions with the road infrastructure, receiving injuries from an impact with the road surface or guard rails.

Among car occupants, the majority (60%) received their lower leg or knee injury from impact with the interior of their vehicle during a collision. In the English data contacts causing the knee or lower leg fracture were from contact with the facia panel, rigid bracketry behind the facia panel, or the footwell. Other injury causation codes recorded for knee and lower leg fractures were from rigid bracketry behind the steering column, the lower A-pillar in the footwell, the pedals, the bulkhead, the back of a front seat or the interior of the side door. The average delta-V in the crash (change in velocity) from vehicle damage for car occupants receiving a MAIS 3+ lower leg or knee fractures was 45kmh. This makes the collisions within the highest 10% of Delta-V results recorded for all collision types. Additionally, for collisions that result in a MAIS 3+ knee or lower leg fracture the damage details recorded for the observable damage to the vehicle structures indicates that as the Delta-V increases cases of ‘severe crush to the vehicle structure with associated intrusion’ and ‘massive impact damage with loss of vehicle integrity’ become more prevalent.

The majority of passenger car collisions that resulted in MAIS 3+ lower leg or knee fractures were to the front of the vehicle and represented directions of force between 11 and 01 ‘o’ clock. Where intrusion into the passenger compartment is seen (present in 54% of the English cases) the degree of this is relatively extensive with the average intrusion measure recorded as 32cm. 50% of the intrusion measures are recorded as over 20cm.

RECOMMENDATIONS

Recommendations for policy makers

In addition to reducing the number of fatalities, road safety policy making should also be aimed at reducing the number of MAIS3+ casualties and their health impacts. In that respect, the following groups are of special relevance as they show a relatively high number of MAIS3+ casualties in relation to fatalities and/or relatively many years lived with disability (YLD, limited to MAIS3+casualties) compared to years of life lost (YLL): cyclists, 0-17 year olds, spinal cord injuries, knee and lower leg fractures and femur shaft fractures. Road safety policy especially aimed at these groups could further reduce health impacts of MAIS3+ casualties.

The in-depth analyses in this report provide more detailed information on the casus of these crashes and therefore provide guidance for the further development of road safety policy. Concerning cyclists, the most relevant crash types are priority error related collisions and single bicycle collisions. Moreover, measures could be focused on reducing distraction and preventing vision obstruction. On the longer term, pro-active vehicle systems that can detect, predict and resolve priority or ‘give-way’ issues before they occur could be effective for reducing the number of MAIS3+ casualties in bicycle – motor vehicle crashes.

Concerning 0 to 17 year old pedestrians, in-depth data indicates that crossing behaviour and crossing location choice could potentially increase the risk of vision or attention issues. It is therefore important to better understand crossing behaviour of young pedestrians, with particular emphasis on crossing point choice, judgement of vehicle speed and vision obscuration. Measures could be aimed at better assisting young pedestrians. Moreover, active vehicle safety systems should be able to deal with (unexpected) crossing behaviour of young pedestrians.

The injuries received by occupants of cars, particularly those in the knee/lower leg and spinal cord groups, indicates that there is still some work to be done in terms of passive safety. The large levels of intrusion seen both in planar collisions (front, side and rear) and in rollovers indicate that vehicle structural strength is still an important topic. Finally, vehicle design might also be further improved to better protect PTW users in collision with passenger cars.
Chapter 5 contains a first suggestion for potential countermeasures, selected from the SafetyCube DSS. It should however be noted that did report did not investigate whether these countermeasures are actually effective for reducing the selected groups of MAIS3+ casualties. More research is needed aimed at designing effective measures to prevent the selected groups of MAIS3+ casualties.

Recommendations for further research

Further research would be useful to investigate differences in MAIS3+/fatality ratios between countries. It is feasible that the differences currently seen are for a large part due to differences in specific types of crashes and specific circumstances between countries. To understand these differences more completely it would be useful to see whether it is possible to derive comparable sets of MAIS3+/fatality ratios for different crash types.

The selected groups of MAIS3+ casualties appear in small samples in the investigated in-depth databases. There is clearly a need for more data on the types and causes of the selected groups of MAIS3+ casualties (0-17 year olds, cyclists, spinal cord injuries and knee/lower leg fractures). There is a risk that existing or future in-depth data collection methodologies and sampling protocols could miss relevant cases involving serious injury, for example by continuing to focus on fatal crashes.

Pagina's
91
Gepubliceerd door
European Commission, Brussels

SWOV-publicatie

Dit is een publicatie van SWOV, of waar SWOV een bijdrage aan heeft geleverd.