Bermongevallen: karakteristieken, ongevalsscenario's en mogelijke interventies

Resultaten van een dieptestudie naar bermongevallen op 60-, 70-, 80- en 100km/uur-wegen


Davidse (red.), R.J.



Met weginspecties, voertuiginspecties en interviews is in deze studie zo veel mogelijk informatie verzameld over bermongevallen. Het gaat om alle ernstige bermongevallen die gebeurd zijn in de periode september 2009 tot en met oktober 2010 in het gebied dat samenvalt met de politieregio’s Haaglanden en Hollands Midden, op 60-, 70-, 80-, en 100km/uur-wegen. Het doel daarvan is inzicht te krijgen in de factoren en omstandigheden die van invloed zijn op het ontstaan en de afloop van bermongevallen. Op basis hiervan kunnen maatregelen worden geselecteerd waarmee vergelijkbare ongevallen in de toekomst kunnen worden voorkomen of waarmee de letselernst van deze ongevallen kan worden teruggedrongen.

Run-off-road crashes: characteristics, crash scenarios and possible interventions; Results of an in-depth study of run-off-road crashes on 60, 70, 80 and 100 km/h roads

This report presents the results of a SWOV in-depth study of run-off-road crashes. In an in-depth study of road crashes all possible information is collected about all aspects of the crash: the traffic conditions, the immediate environment, the road users who are involved, their vehicles, and the injuries that have been sustained by the vehicle occupants. The purpose of the present research is to gain insight in the factors and circumstances that have an influence on the occurrence and the outcome of run-off-road crashes. This will provide a basis for the selection of measures that can prevent similar crashes in future or reduce the injury severity of these crashes. This is advisable because one third of all fatal road crashes are run-off-road-crashes. Furthermore, the number of fatal run-off-road crashes does not decline as rapidly as the total number of fatal crashes.

Run-off-road crashes are defined as crashes in which one of the motor vehicles that are involved runs off the road in the initial phase of the crash. This in-depth study limits itself to those run-off-road crashes in which a driver hit the verge. The end position of the vehicle is not important; the vehicle may have come to a stop on the verge, in a ditch, against an obstacle or tree, but it may also have crashed into an approaching vehicle. Therefore, run-off-road crashes are not only single-vehicle crashes or obstacle crashes; they can also be frontal or lateral collisions provided that one of the involved vehicles ran off the road in the initial phase of the crash.

Data gathering

The in-depth study into run-off-road crashes was carried out in two Dutch police regions: Haaglanden and Hollands Midden. In this area all run-off-road crashes involving drivers were selected that had occurred on 60, 70, 80 and 100 km/h roads during the period 1 September 2009 to 31 October 2010. These formed a total of 28 run-off-road crashes. The majority of the run-off-road crashes, 25 out of 28 (89%) were single-vehicle crashes; the driver who hit the verge had not been in contact with another road user. In three of the crashes (11%) an oncoming vehicle was hit when the driver attempted to steer the vehicle back onto the road. Although the driver of the oncoming vehicle did see that the other driver had trouble regaining control of his vehicle, it was too late to swerve and avoid a crash.

For all 28 run-off-road crashes the SWOV research team for in-depth studies gathered information about the drivers who were involved (using interviews), about the traffic conditions (using road inspections), about the damage to the vehicles (using vehicle inspections) and/or about the injury that may have been sustained  by the vehicle occupants (using interviews and supplementary data files).

The interviews were conducted by a psychologist on the research team. Of the 31 drivers, 35% were willing to take part in the study.  Of the 28 drivers who had hit the verge, 32%, all of them male, were willing to take part.  In addition, one co-driver took part in the study instead of the driver. Of the 22 male drivers who had hit the verge, 41% participated in the study through an interview or questionnaire (see Table 1). Not one of the six female drivers was willing to participate in the study. This means that less information is available about the crashes they were involved in; this may slightly affect the findings of the present study.


Drivers cooperating





9 (41%)

 0 (  0%)





60 or older


 3 (50%)

 2 (17%)

2 (33%)

2 (50%)

Table 1. Number and percentage of drivers who had hit the verge and who participated in the study by means of an interview or questionnaire.


During the interview, questions were also asked about the injuries that had been sustained by the occupants. In addition, their permission was asked to retrieve medical information about the occupants’ injuries from the hospital. Both the self-reported injuries and the medical data were used to determine the injury severity.

The inspections of the crash locations were carried out by a traffic engineer on the research team, supported by a project assistant. All of the crash locations were inspected. The involved vehicles were also inspected.  This could be done for 25 of the 31 vehicles (81%). The other vehicles were no longer available. All involved vehicles were passenger cars.

Approximately half of the studied run-off-road crashes (54%) occurred on an 80 km/h road. This percentage agrees with the proportion of 80 km/h roads in the total road length in the Netherlands (see Table 2). However, in relation with the share of 80 km/h roads in the area that was studied, the proportion of run-off-road crashes is about twice as large. This indicates that the composition of the road network in the area studied is not representative for the Netherlands: the area has relatively few  80 km/h roads and relatively many 60 km/h roads. On the 80 km/h roads in the area, however, relatively many run-off-road crashes occurred. In comparison with the road length, relatively few run-off-road crashes happened on the 60 km/h roads.


Studied run-off-road crashes


Rural road length


All serious crashes in rural areas involving a passenger car (2007-2009)





60 km/h roads






70 km/h roads






80 km/h roads






100 km/h roads






Other roads (including 120 km/h roads)

Not included in the selection











Table 2. Proportions of the rural road types as location for run-off-road crashes in rural areas, in the total road length and in the total number of registered serious crashes in rural areas involving passenger cars.

Characteristics of run-off-road crashes

When the entire group of 28 run-off-road crashes is considered, these crashes are often seen to occur in the weekend, on an 80 km/h road, and in a bend (see Table 3). This is shown by comparison with the distribution of all serious crashes on rural roads in the Netherlands in which at least one passenger car was involved. The driver running off the road is often male (79%). Furthermore, as a group young men (aged 18-24) are overrepresented among drivers running off the road. Another group of drivers who are relatively often involved in run-off-road crashes are men aged 30 - 39.


Proportion (%) in the number of run-off-road crashes (N=28)


Male (79%)

Age driver

18-24 (21%), 30-39 (29%), young males (21%)

Day and time of day

Weekend (57%)

Road type

80 km/h roads (54%)

Road characteristic

Bends (48%)

Table 3. Most frequent characteristics of run-off-road crashes.

Of the 28 drivers who were involved in run-off-road crashes, 18 (64%) were checked for alcohol. One of them appeared to be above the legal level (a blood alcohol concentration of more than 0.5 ‰).

Subtypes of run-off-road crashes

The crash process was analysed in more detail for 27 of the 28 crashes, as there was not sufficient information to determine at which location one of the crashes had occurred. Because an interview was not present either, this crash was excluded from further analysis. For each of the 27 crashes that were studied in detail, the in-depth research team tried to determine the course of events and which factors had played a role in the occurrence of the crash and possible injuries. A distinction was made between factors related to the driver of the vehicle, to the vehicle itself, to the road, and to general conditions at the time of the crash. All relevant factors were selected, as the starting point of the analysis was the assumption that a crash is the consequence of a confluence of events and that multiple factors play a role in the causation of crash and injury.

After all 27 crashes had been described in this manner, the crashes with similar crash processes (comparable causation and combination of contributory factors) were grouped into types of run-off-road crashes. Next, these types of run-off-road crashes were described based on the circumstances in which the crashes had taken place, the drivers who were involved and the contributory crash factors. A total of four types of run-off-road crashes were identified, leaving a remainder of six crashes. Too little information was available about these crashes to allow categorisation.

In Table 4, the characteristics of these four types of run-off-road crashes have been summarized. The middle column gives a description of the prototypical scenario for each subtype. This scenario contains the greatest common factor of all crashes of that specific subtype. This means that it is not a real crash, but a description by characteristics of that specific type of run-off-road crash.

The contributory crash factors that are given in the right-hand column of Table 4 are the result of the deliberation of every single crash by the SWOV in-depth research team. During these deliberations, all relevant factors were selected that according to the team had contributed to the occurrence of the crash and the injury of the occupants of the vehicles involved. To evaluate the factors relating to the road, the characteristics of the cross-sectional profile were compared with the guidelines that have been drawn up by CROW, the Dutch Information and Technology Platform for Transport, Infrastructure and Public Space. Expressions like ‘too narrow’ and ‘too steep’ are the result of such comparisons. A departure from the guidelines was not necessarily labelled as being ‘wrong’; it is never automatically a contributory crash factor. This was dependent on the entire course of the crash. Similarly, the fact that someone is a novice driver and holds a beginner’s licence, is not sufficient reason to select the beginner’s licence as one of the contributory factors. Driving behaviour and/or vehicle control must also give reason to do so. However, there was not always enough evidence to support this. If there was reason to assume that a certain factor had played a role in the crash, but evidence was not entirely conclusive, doubts about the validity of that factor were reported. In Table 4 this is shown by the margins that are given behind the contributory crash factors. The first (and lowest) figure indicates in which percentage of the crashes the contributory crash factor (almost) certainly played a role. The second percentage also includes those crashes in which there was some doubt about the validity of that specific factor. 


(number of crashes and their share of the total number of analysed run-off-road crashes)

Description of the prototypical scenario

Most frequent crash factors*

Risky driving behaviour

(n=7; 26%)

A young novice male driver (18-24 years old) is driving on a weekend night, under favourable road and weather conditions, with age-peer passengers. While driving he takes a risk by overtaking at a location where this actually is dangerous, or by taking a bend at high speed.  This risk-taking may be intentional, but it can also be the result of overestimation of one’s driving skills. The high driving speed while taking a bend could be caused by a warning that is lacking and/or guidance in a tight bend. Next, the driver loses control of the vehicle and crashes into an object that is situated within the obstacle-free zone that is required for a road with that particular speed limit. As a consequence of the crash the driver or his occupant sustains fatal injury (with a MAIS of 4 to 6) or sustains minor to moderate injury (MAIS of 1 or 2).

  • Lack of experience (57-71%)
  • Driving speed too high (57-71%)
  • Bend not well-announced (43-57%)
  • Obstacle-free zone too narrow (57%)

Temporary inability to react

(n=7; 26%)

A young (under 40) or, on the contrary, older (65+) driver is driving a passenger car on a weekend day during daylight and under favourable road and weather conditions. While driving he is struck by a black-out or he falls asleep. The driver  – in most cases a single occupant  – does not regain consciousness or wake up until he hits the verge or an obstacle.  In the first case he still makes an attempt to get the vehicle back on the road, but fails to regain control. The vehicle comes to a standstill against a tree that is situated within the obstacle-free zone that is required for a road with that particular speed limit. The risk that the driver hits the verge in the initial phase of the crash increases because the road surface is too narrow and or because the straight road changes into a bend. As a consequence of the crash the driver is not injured or sustains minor to moderate injury (MAIS of 1 or 2).

  • Fatigue (29-57%)
  • Black-out (29-57%)
  • Lane too narrow (0-43%)
  • Obstacle-free zone too narrow (43-57%)


(n=2; 7%)

A driver, who is driving on a straight road during daylight and under favourable road and weather conditions is distracted by something that happens at the side of the road and has nothing to do with traffic. This causes a gradual deflection from the course and the vehicle hits the verge. The driver tries to steer the vehicle back onto the road, but hits an object and/or oncoming vehicle which causes his vehicle to roll over. As a consequence of the crash the occupants are not injured or sustain minor injury (MAIS 1).

  • Distraction (50-100%)


(n=5; 19%)

A young male or female driver (under 40 years of age), the sole occupant of the car, is driving on a weekday under favourable road and weather conditions. While driving he or she is frightened by an unexpected occurrence in traffic. This causes the driver to suddenly change course. The unexpected factor can be an oncoming vehicle or an animal in the road, but can also be in the form of being blinded by an oncoming vehicle or the sun. The risk of being frightened followed by a sudden course change may in some cases be increased by fatigue or being deep in thought. After the sudden steering, the vehicle hits the verge and in many cases ends up in a gulley (water channel). As a consequence of the crash, the occupants are not injured or sustain minor injury (MAIS 1). An exception is a crash in which the vehicle landed in a river upside down; this resulted in a MAIS of 4.

  • Human factors unknown (60%)
  • Other road user (20-80%)
  • Glaring (0-40%)
  • Tyre (20-40%)

* The first (and lowest) figure in brackets indicates the percentage of crashes in which the contributory crash factor (almost) certainly played a role. The second percentage also includes those crashes about which there was some doubt about the validity of that particular factor.

Table 4. Summary of the subtypes of run-off-road crashes. The percentages refer to the proportion in the subtype in question.

When Table 4 is studied, a number of patterns can be distinguished.

Risky driving behaviour: in the weekend having friends in the car

Young male drivers are particularly involved in run-off-road crashes which occur after risky driving behaviour (subtype 1). In comparison with the other types of run-off-road crashes these crashes relatively often occur in the weekend (71%) and in the dark (71%) and the driver relatively often has one or more passengers in his car (71%). Contrary to the picture that is sometimes drawn, alcohol did not play a role in these crashes occurring.

Falling asleep after a busy working week

The run-off-road crashes that occurred after a driver was temporarily unable to react due to falling asleep or a sudden medical condition (subtype 2), also frequently occur in the weekend (71%), but during daylight (86%) and only in 29% of the cases were one or more passengers present in the car. The falling asleep seems to be related more with work (night shift, end of a busy week) than with going out.

Swerving to avoid something or someone else

The run-off-road crashes that occur after the driver has changed course due to an external factor (subtype 4) often happen on a weekday (80%). The drivers who were involved in these crashes are known to have changed course to avoid another road user or an animal. However, there is little information about their behaviour or mental condition previous to the course change. The explanation can be found in the limited cooperation in the study by the drivers involved. Therefore, one of the matters that are unclear is whether driving experience was a factor in hitting the verge.

Run-off-road crashes in bends: risk taking or falling asleep 

Just over half of the run-off-road crashes that occur after a driver hits the verge due to risky behaviour (subtype 1), happens in a bend (57%). In the majority of this type of run-off-road crashes (71%) too high a driving speed played a role in the occurrence of the crash. At the same time, the speed limit in force is relatively low  (86% lower than 80 km/h). In this subtype, the speeding in bends (four crashes) is probably partly related with poor warning and guidance in the bends. A warning for and/or delineation of these bends was necessary because the radius was too tight, even at the low local speed limit of 50 km/h.

Also for the crash type ‘temporary inability to react’ (subtype 2) half of the crashes occurred in a bend (43%). In these crashes, however, poor warning and/or delineation of the bend did not play a role. The drivers who ran off the road due to ‘loss of consciousness’ probably did not notice the bend at all and went straight ahead instead. The drivers who ‘took a risk’, on the other hand, lost control in the bend because they made an incorrect assessment of the bend and therefore approached the bend at too high a speed and/or because the radius was too tight for the local speed limit.

Serious consequences

Regarding the seriousness of the crashes it is striking that the crash types ‘risky driving behaviour’ (three fatal crashes) and ‘changing course’ (one MAIS 4) have the most serious consequences. In the latter of these two types (‘changing course’) the severity is especially influenced by one crash in which the car landed in water upside down. In the other type (‘risky driving behaviour’), a similar situation also ended very seriously. In the remaining very serious crashes of the type ‘risky driving behaviour’, the severity of the consequences was mainly determined by a not collision-friendly object within the obstacle-free zone that is desirable for the local speed limit.

Vehicle does or does not roll over

Regarding the end position of the vehicles, it is striking that half the vehicles rolled over or ended upside down in crashes of the type ‘risky driving behaviour’, whereas this did not happen on any single occasion in crashes of the type ‘temporary inability to react’. On the basis of the other differences between the characteristics of these subtypes several explanations are possible, such as the high driving speed and the driver being inexperienced on the one hand, and failing to act on the other. However, the actual reason (or combination of reasons) is not easily established.

Contributory crash factors of run-off-road crashes in general

For each category of contributory factors (general, human, vehicle and road), Table 5 indicates which factors most frequently played a role in the total set of 27 analysed run-off-road crashes, irrespective of subtype.

Type of factor

Most frequent contributory crash factors

(% of total number of  27 analysed crashes)a

General factors


Wet road surface (4-15%)

Darkness (4-11%)

Human factors


Distraction (19-30%)

Driving speed too high (15-19%)

Novice driver (11%)

Fatigue (7-19%)

Vehicle factors

Tyres (3-11%)

Road factors


Obstacle-free zone too narrow (44-52%)

Slope too steep (22-26%)

No semi-hard shoulder (19-26%)

Traffic lane and/or hard shoulder too narrow (11-26%)

Bend: curve radius too tight and not announced or delineated correctly (11-15%)

a The first (and lowest) number in brackets indicates the percentage of crashes in which the contributory crash factor (almost) certainly played a role. The second percentage also includes those crashes about which there was some doubt about the validity of that particular factor.

Table 5. Summary of the most frequent contributory factors in crashes.

Some factors co-occur relatively frequently. A too narrow obstacle-free zone and a too steep slope is the most prominent combination. In 50% of the twelve run-off-road crashes in which a too narrow obstacle-free zone (most) probably played a role in the occurrence and consequences of the crash, too steep a slope (steeper than 1:3) was also considered to be a contributory crash factor. In 67% of the cases this slope led straight to a water channel (four crashes). In two of these four crashes this had very serious consequences (MAIS 4 or 5) due to (near) drowning.

Injury and injury factors

In the 27 run-off-road crashes that were analysed in detail, 30 vehicles were involved with a total of 52 occupants. Of these occupants, 8% sustained fatal injury and 15% had a MAIS of at least 2 (serious road injury). The remaining occupants who had sustained injury, 29%, (probably) had a MAIS of 1. The group of occupants without injury was about the same size, 31%. Of 17% of the occupants the injury sustained in the crash was unknown.

The most frequent injury factors have been summarized in Table 6.

Injury factor

Percentage of the 52 vehicle occupants for whom this factor played a role *

Contact with the interior of the vehicle


Vehicle rolling over/upside down


Speed before impact


Injury caused by safety devices

seat belt: 10%

airbag:   8%

* More than one factor can be attributed to an occupant.

Table 6. Most frequent injury factors.

The most serious injury (5 occupants with a MAIS of 4 or higher, including 4 road fatalities) was caused by contact with obstacles that are not collision-friendly (trees and a lighting pole without breakaway or shear off construction) and the vehicle landing in water upside down. The obstacles mentioned above were not fenced off and were located within the desired obstacle-free zone given the speed limit on the roads concerned (CROW, 2004a). The water, which was not fenced off either, was even located within the minimally required obstacle-free zone.

The obstacle-free zone is intended to give the road user the opportunity to come to a safe stop at the current speed limit. This zone should not contain obstacles that can cause serious damage to a vehicle and/or injury to the occupants (CROW, 2004a). In 15 of the 27 run-off-road crashes at least one not collision-friendly object was present within the desired obstacle-free zone (see Table 7).

The presence of a not collision-friendly object within the obstacle-free zone was partly responsible for the injury of 19 of the 52 occupants (37%) involved. The crash into the object led, for example, to contact with the interior of the vehicle (e.g. door or windscreen), which then resulted in injury (see Table 6 above).


Number and percentage of vehicles



7 (23%)


5 (17%)

Slope (without gulley/ditch)

2 (7%)

Lighting pole

1 (3%)


15 (50%)

Table 7. Obstacles within the desired obstacle-free zone against which a vehicle has come to a stop or which have played a role in sustaining injury. The proportion is expressed as the percentage of the vehicles that were involved in the run-off-road crashes that were analysed.

The vehicle rolling over also caused contact with the interior of the vehicle. A total of 11 vehicles (37%) rolled over during the crash. Of the total number of occupants, 21 (40%) travelled in a vehicle that rolled over or turned upside down. Twelve of these occupants are known to have sustained injury in the crash. For ten of them (19% of the total) their injury was (partly) attributed to the rolling over or turning upside down of the vehicle in which they travelled.

The use of safety devices can prevent injury or reduce its severity. Twenty-nine of the occupants (56%) are known to have worn a seat belt. The fact that a seat belt was used (most) probably contributed to the reduction of injury for 26 occupants (50%), and for ten of them the airbag also contributed. Five occupants (10%) did not use a seat belt or were unable to use it because it was not present. For 19 occupants it is unknown whether or not a seat belt was worn.

It is easier to establish whether airbags were operational. 43 of the 52 occupants travelled in one of the front seats of the vehicle. Thirty of these front seats (70%) were fitted with one or more airbags. Less than half of the airbags (43%) unfolded during the crash. However, the percentages of airbags that unfolded differ largely between the types of run-off-road crashes: from 13% in the group of ‘others’ to 88% in the group of run-off-road crashes that occurred because a driver was unable to react due to fatigue or a black-out. Other differences between these two subtypes provide a basis for two possible explanations for the variance in the percentages of airbags that unfolded:

  1. In the latter subtype of run-off-road crashes the majority of the vehicles crashed into a tree, whereas in the other subtypes more of the vehicles rolled over and came to a standstill in a gulley (water channel) or at the bottom of a slope. Front airbags are mainly intended to absorb the impact force in frontal crashes and will therefore be more readily activated in a crash against a tree than in a vehicle rolling over.
  2. In the latter subtype of run-off-road crashes the driver has had little or no time to act and therefore the crash generally occurred at a higher speed.

Comparison with findings in other studies

The results of the present in-depth study have been compared with the results of other in-depth studies. Most attention was paid to a study which, using the same methodology, was carried out parallel to the present in-depth study in another, more rural region in the Netherlands (see Davidse et al., 2011). That study, which was carried out in the Province of Zeeland, found the same injury and contributory crash factors as were found in the present study. At the same time, the Zeeland in-depth study provided deeper insight. The greater number of run-off-road crashes that could be analysed and the greater number of interviews that were available allowed the researchers to make a more detailed subdivision  of the types of run-off-road crashes.  In addition, the study deepened the understanding of the human-related contributory crash factors. The knowledge from this parallel study in Zeeland has been utilized in the process of selecting measures that can be used to break through the crash patterns that have been identified. Therefore, the measures that are discussed in this report are in keeping with the crash and injury factors that have been identified in the present study and/or in the parallel Zeeland study.

Measures to prevent run-off-road crashes and/or reduce the severity of the outcome

The previous sections indicated that an obstacle-free zone that is too narrow plays a role in the occurrence of about 40% of the run-off-road crashes. A not collision-friendly obstacle within the 'obstacle-free' zone created a 'danger zone', and deprived the driver from coming to a safe standstill. An important measure to prevent run-off-road crashes and to reduce injury severity is, therefore, moving or fencing off obstacles that are positioned in the obstacle-free zone. This means that priority should be given to following the guidelines with regard to obstacle-free zones, as described in the CROW Handbook for the safe layout of roadsides (CROW, 2004a). The implementation of other infrastructural measures described in this handbook will also improve safety. Based on the contributory crash factors that were discussed in the present in-depth study and the Zeeland in-depth study, the following five measures from the handbook are expected to be the most effective in reducing the number of run-off-road crashes (compared with the other measures that are presented in the handbook). The measures are listed in the order used by the handbook, which is not an indication of effectiveness:

  • implement speed enforcement and/or lower the speed limit;
  • apply a profiled, acoustic edge marking or install rumble strips on the hard shoulder;
  • move obstacles to outside the emergency and recovery zone, and preferably also outside the minimally required obstacle-free zone, or remove these obstacles altogether;
  • make slopes, ditches, and banks of gulleys more gentle, and apply round tops and foots;
  • apply safety barriers that lead to lower injury severities than driving into the danger zone.

Each of these measures relates to a contributory crash factor that plays a role in 20 to 40% of the run-off-road crashes that were studied in the present in-depth study and in the parallel Zeeland study. For the moving of obstacles it must be noted that in the above-mentioned in-depth studies, the team had the desired obstacle-free zone in mind when selecting the obstacle-free zone as a contributory factor, and not the minimally required obstacle-free zone. To get anywhere close to the above-mentioned 40% of the total number of run-off-road crashes, the objects will therefore need to be placed further from the pavement than the width of the emergency and recovery zone, and also further than the minimally required obstacle-free zone for the design speed at that location. This is advisable for all roads and not only for those roads where crashes occurred. The latter advise also pertains to the other measures.

In addition to the infrastructural measures that are discussed in the CROW Handbook for the safe layout of roadsides, a number of ‘new’ measures has been selected which, in the opinion of the in-depth team, fit the combinations of contributory factors that were identified in the present study and in the parallel Zeeland study. These supplementary measures were selected in a brainstorm session with SWOV experts on different disciplines (infrastructure, vehicle, human behaviour). One of the supplementary measures mainly focuses on the road layout (road factor) and is an elaboration of two measures from the above-mentioned CROW handbook (horizontal alignment and/or improving the delineation of tight bends). An important part of the bends in which run-off-road crashes had occurred (46% in the research area of the present study and 86% in the province of Zeeland), had a bend radius that was too tight for the speed limit (assuming a standard superelevation of 2.5%). Such bend radii must be delineated in accordance with the CROW guideline for the marking and signposting of roads. An inspection of the approach roads indicated that 83% and 88% respectively of the tight bend radii were not delineated in accordance with the guidelines. An inspection of tight bends and – wherever necessary – adaptation of the layout and/or marking and signposting of these bends is therefore a useful measure to reduce the number of run-off-road crashes.

The contributory factors which relate to human behaviour and were most frequently identified in the present study and in the parallel Zeeland study are distraction (19% and 31%), speeding (15% and 27%), fatigue (7% and 17%), young novice driver (11% and 10%) and alcohol use (4% and 19%). In all cases, poor state awareness plays an important role. State awareness is about ‘knowing what you are capable of doing, knowing the dangers of certain behaviour or traffic situations and adjusting your behaviour to allow safe traffic participation’. The state awareness of road users can be improved by public information and education, and by making use of in-vehicle information systems that provide feedback about traffic behaviour or warnings about a slippery road surface or unexpected traffic conditions like a traffic jam or roadworks.

Other measures aimed at vehicle and infrastructure that are expected to reduce the influence of the above-mentioned contributory factors which are related to human behaviour are:

  • transverse ridges on the traffic lane on the approach of a bend (distraction, fatigue and/or speeding);
  • acoustic or haptic signals inside the vehicle that warn when the driving speed is too high for the bend one is approaching (distraction, fatigue and/or speeding);
  • monitoring of the driver’s condition (fatigue);
  • plan of attack for narrow roads (distraction); and
  • young drivers-ISA (speeding).

A more detailed description of these measures can be found in Chapter 5.

A number of the above measures were selected based on the fact that certain human and road related factors or road characteristics often co-occur. Examples are distraction in combination with crashes in bends (Zeeland in-depth study), high speeds and crashes in bends (present in-depth study and Zeeland in-depth study), and distraction in combination with a speed limit that is higher than what is suitable for that road type, cross section, and the width of the obstacle-free zone (Zeeland in-depth study). This latter combination indicates that run-off-road crashes in which distraction played a role, can also be prevented by using speed limits that are in accordance with the width of the cross section and the accompanying obstacle-free zone. In the present report, this is referred to as the ‘plan of attack for narrow roads’. A suitable speed limit informs the driver about the road user behaviour that is suitable for the layout of road and roadside. At this speed, drivers will have sufficient time to correct for a deviation from their course. Public information can be used to enforce the correct driving speed. This public information is also part of the proposed ‘plan of attack for narrow roads’.

To reduce injury as a consequence of running off the road, and rollovers the following supplementary measures are advised: 

  • cover gulleys and ditches with a cattle grid or use different means to level them out;
  • make airbag and seatbelt (even) more intelligent;
  • prevent abrupt steering with a vehicle system that is yet to be developed and stimulate use of the already developed electronic stability control.

If all the above measures are classified by the phase of the run-off-road crash they are relevant for, this results in the set of measures that is shown in Table 8. Part of these measures are taken from the CROW Handbook for the safe layout of roadsides (CROW, 2004a) and concern the implementation of existing guidelines (in Table 8 indicated by the text ‘[Guideline]’). Compliance with the guidelines can be improved by using audits and road safety inspections. Preferably these instruments will be incorporated in a quality assurance system. The other measures are of a more innovative nature (‘[Innovation]’). Further research into their feasibility and effectiveness will need to be carried out before these measures can be implemented.



Print this page