3. ROAD SAFETY REVIEW PRINCIPLES
An overview of the road safety review process
3.1 Preamble
To convey our message clearly, we need to state at the outset what we believe to be true about safety in
road design.
3.2 There are no safe roads
A road would be completely safe if no collisions occurred on it. But crashes occur on all roads in use.
It is therefore inappropriate to say of any road that it is completely safe. However, it is correct to say that
roads can be built safer or less safe. Consider two alternative road designs, connecting the same two
points and carrying the same traffic. The road design that is likely to have fewer or less severe crashes
would be deemed to be the safer one.
3.3 The nature of safety improvements
We know how to make a road safer. Building a wider median, placing obstacles farther from the travelled
lanes, providing more pavement friction, designing curves with larger radii, etc., all make for safer roads.
Typically, safety changes gradually as some dimension of the road or of the roadside changes. Therefore,
it makes little sense to say, for example, that if obstacles are 9 metres (m) from the edge of the travelled
lane the road is unsafe, but if they are 10m from it, the road is safe. What is true is that the farther an
obstacle is from the edge of the travelled lane, the fewer and less severe collisions with it are likely to be.
Some safety improvements are not gradual. Thus, for example, the decision to illuminate the road will
cause an abrupt drop in nighttime collisions and a (usually smaller) increase in daytime collisions with light
poles or the barriers in front of them.
3.4 Diminishing marginal benefits
Safety improvements are usually subject to the law of diminishing marginal returns. This means that for
every improvement of a fixed amount, the safety benefit gained decreases a little each time. For example,
increasing the width of the median from 50m to 60m will decrease the number of collisions less than
increasing it from 10m to 20m. Eventually, a width will be reached at which widening the median further
cannot be justified because the improvement in safety is too small.
3.5 A possible objection?
Some people may object to the judgment that a point exists beyond which further improvement in
safety is not justified, claiming that any improvement in safety is worthy. This position, we think, is not
tenable. Expenditure of public money can save life and limb in many places in Ontario. However, spending
the budget where it saves few lives means that some lives will be unjustifiably lost by not spending it where
more lives could have been saved.
3.6 Knowledge, judgment and standards
Knowledge of how the features of a road affect safety is imperfect. Some road features are well-researched
and their effect on safety can be expressed by numbers. For other, less well-researched features, only the
direction of their effect on safety is known. There are even road features the effect of which on safety is at
present unclear. It must be understood that the relationship between road features and safety is not known
with the kind of precision that is customary in the physical sciences and in engineering disciplines that
allow experimentation.
As a result, one must leave room for judgment and legitimate differences of opinion. One must not think that
it is always possible to determine by calculation the point beyond which further improvement in safety is not
cost-effective. The hope is that the collective professional judgment about safety, economics and other
matters is reflected in the engineering standards and practices used in road design. Indeed, road design
standards and practices are the product of accumulated experience and written by committees of experts
in road design. As such, road design standards capture what is agreed upon to be good practice by the
members of the committee at the time the standard is written.
3.7 Road design standards and safety
At this point, a difficult problem of communication arises. In English, "meeting a standard" is understood to
be a guarantee of quality; conversely, if something is "substandard", it is understood to be deficient. This meaning does not apply to safety in road design standards. To explain why, three essential
aspects of the complex relationship between road safety and road design standards are discussed below.
First, as mentioned earlier, the safety of a road does not change abruptly when some road feature changes
slightly. Consider, for example, the current design standard stipulating that the length of an acceleration
lane (see Figure 3.1) has to be at least 500m.
An acceleration lane that is 490m long is not unsafe; it will be, perhaps, only slightly less safe than a
500m-long acceleration lane. Thus, not meeting a standard does not necessarily make a road unsafe.
Second, many highway design standards are limit standards, which means that for a certain class of road,
the radius of a horizontal curve has to be at least "x" metres long, a roadside obstacle must be at least
"y" metres from the edge of the outer lane, the grade must be at most "z" per cent, and so on. Just
meeting such a standard does not make the road as safe as it can be; if a radius larger than "x" is chosen,
if obstacles are placed farther than "y" from the travelled lane and if the road grade is less than "z", the
road would be safer. Thus, the aim of good design practice is to exceed the limit standard, not just to
meet it.
Third, road design standards evolve with time. Roads used to be made with lanes 3.6m (12 feet) wide;
now the standard calls for 3.75m-wide lanes. This does not mean that the entire old stock of roads with
3.6m lanes is unsafe. Similarly, crest vertical curves over hills used to be designed so that the driver could
see in time to stop safely before hitting an obstacle more than 150 millimetres (mm) high in the path of the
vehicle. Nowadays, the standards say that only obstacles 380mm or higher need to be seen in time.
Thus, by the current standard, obstacles less than 380mm high in the path of a vehicle may not be
visible in time for a safe stop. This does not mean that roads designed with the 380mm obstacle in mind
are unsafe. It means only that the information, judgments and economic considerations that go into
formulating design standards change over time.
Road design standards are not the demarcation line between safe and unsafe. They are a reflection of
what a committee of professionals of that time considers to be good practice.
3.8 Prevention, not fault
Consider the following invented sequence of events. A driver was going north on an arterial road at 70
km/h, where the speed limit is 50 km/h. He intended to take the inner loop ramp
(see Figure 3.2) onto a freeway to travel west. This ramp has a tight curve,
turning right, that has a 50 km/h design speed. The posted advisory speed on the curve is 30 km/h.
The driver apparently misperceived how much deceleration was needed to negotiate the curve and the
vehicle skidded to the left. Since there is no guiderail at this point, the vehicle rolled down the
embankment. The embankment is a drop of 5m over a horizontal distance of 15m.
The rear door opened, an unbelted child was ejected and severely injured.
Clearly, the driver was "going too fast for the conditions" and the child should have been wearing a seat
belt. But fault or culpability is not the issue here. For the highway designer, the question would be:
"What might help prevent collisions of this kind or reduce their severity?" If we could make drivers
go slower at this site and help them perceive what deceleration is needed, and if the curve were built
less tight, (i.e. had a higher design speed been chosen), fewer vehicles would run off the road
on this ramp. If a guiderail were placed along the entire curve, some skidding vehicles would be
prevented from tumbling down the embankment. Even if there were no guiderail, but the
embankment had been built to be less steep, some vehicles might not overturn going down the slope.
If vehicles were made so that doors were less likely to open when the vehicle rolled, there would be
fewer ejections of occupants. If we could induce more occupants to wear seat belts, this, too, would
reduce the chance of ejection and injury. All of these and several other actions might have altered
the course of events and the final outcome. However, only some of these actions would be within
the direct purview of the highway designers. At this site, the designer should examine at least:
- the choice of the design speed (and the resulting minimum radius of the ramp curve);
- the decision of where guiderails would be placed; and
- the design of embankment slopes.
Obviously, the driver's "human error" played a major role in this invented story, as it does in most real
collisions. This leads many to think that road users ought to be the sole target of preventive measures.
Among road safety professionals, however, such thinking is widely recognized as incorrect. The fact
that almost all collisions could have been prevented had those involved acted differently does not
mean that the most effective way to reduce collisions is to alter people's behaviour or tendency to
make errors. Effective action must aim jointly at the human element, the vehicle and the road. Road
design can reduce the incidence of human error and the chance that human error will result in a collision,
and can ameliorate the severity of the consequences of collisions that are initiated by human error.
3.9 Comments and questions
In this chapter, we have described our view of road safety in highway design. In summary, we think that
roads can be built safer or less safe, but that a road cannot be classified simply as being either safe or
unsafe. Meeting minimum standards does not guarantee that a road is sufficiently safe. Conversely, not
meeting a minimum standard does not necessarily imply that the road is unsafe.
We often know the direction in which safety is likely to change if some road design feature is altered.
However, current knowledge about the amount of safety gain or loss that comes from a change in some
road feature is imprecise. Also, the design choices engineers make imply trade-offs between life, limb and
economics, a judgment that is at best difficult.
In view of these beliefs and the circumstance of this review, we chose to answer the following questions.
- Does Highway 407, as built, meet or exceed the prevailing Ontario design standards
that have a bearing on road safety?
This question is raised because the public debate has been couched in similar terms and therefore an
answer will be expected.
- Were the standards used and the design decisions made in a manner that appropriately
addresses safety?
- Were cost-effective opportunities to enhance the safety of Highway 407 examined
and exploited during its design?
We chose to raise these questions because meeting minimum standards does not mean that the road
is sufficiently safe and not meeting minimum standards does not necessarily imply important safety omissions.
- What seemingly cost-effective opportunities to enhance the safety of
Highway 407 exist now?
This question must be answered to ensure that no reasonable measures to enhance public safety are
overlooked.
Figure 3.1. Freeway acceleration lane 