Road design, safety and Vision Zero

[plam]

It’s such a weird view.  Even if we keep narrowing down what you’re talking about, it’s still easy to show that you’re wrong.  Bridge designers know that there’s some probability that someone will jump off their bridge and kill themselves.  In aggregate, they absolutely know for sure that people are dying from their designs. And yet, we don’t strive to make it impossible* for someone to commit suicide.

It’s the same with road designs.  Any given stretch of road has some probability of a fatality.  In aggregate, it’s a guarantee someone will die.


* note: we almost certainly don’t do as much as we should.  So I’m not saying we shouldn’t do more - even while still saying we should accept we can never totally prevent bridge suicides.

But we do retrofit pedestrian bridges and build new bridges so that it's quite hard to jump off, with overhanging fences. (Interestingly, this intervention seems to decrease the overall suicide rate as well; when people are deterred, it seems that surprisingly often they just decide not to go through with it after all.)

So the carnage rate on our highways is relatively low and has been decreasing, but it's still quite high because people spend a lot of time on the highways. We do try to make efforts to reduce this rate, but I do tend to agree that we're not trying really hard. As an example of trying hard in another domain, I was reading about how Amazon was working on improving shipping timeliness before they rolled out Prime. Every time they missed their deadline they would figure out how not to do that again. We just don't have that level of effort for reducing collision rates.

[jamincan]

There are loads of improvements engineers have made to highways over time that improve safety. Tall wall barriers undoubtedly do that. Parclo interchanges were in part designed to reduce weaving of traffic.

Two problems that still seem pervasive, though. First, it is difficult to test improvements rigorously to see if the improvements actually improve safety. Wider lanes seem safer, but they also increase speeding and can make urban roadways far less safe for other users. Second, engineers still seem to treat other road users as an after thought and design roadways for drivers first, and then figure out how pedestrians and cyclists can fit in after.

[danbrotherston]

And, of course, this idea that road fatalities are just accepted and never investigated is also ridiculous.  Just because we don’t have the resources to investigate every accident in detail like we do for major plane accidents doesn’t mean it never happens.

Major accidents (like say the Humboldt Broncos accident) are looked at in depth and recommendations are made to try to prevent similar accidents from ever happening again.  Wide scale accident statistics are looked at and acted on.  For example, we have concrete barriers for most of our divided highways because we found a common cause of death was crossing over through the ditch into oncoming traffic.

I did not say "never", I said, when the number does not exceed the expected number.  It's literally exactly what the city does.  The region has a list of intersections that they investigate for collisions.  That list is ordered by how much the injury rate EXCEEDS THE EXPECTED NUMBER.  I.e., how many more people are killed beyond the number they expect to be killed--you know, the number that it's okay to kill or injure.

Go ask a traffic engineer, they have actual numbers for the expected number of injuries and deaths for different intersections in the city, including brand new ones.  This is how the field works.

The situations you describe are when there was "more than expected" collisions.  I am not joking here, I'm not exaggerating, I am being absolutely serious. Traffic engineers have an rate of injury and death that is considered normal, and only in cases where that rate is exceeded is there a "problem".

Tou think that all engineering disciplines are the same, that isn't true, no bridge designer has an acceptable number of collapses greater than zero.  Yes, sometimes bridge do collapse, due to mistakes, or lack of repair, or misunderstanding the design, but those are considered failures.  Yes, sometimes there are suicides, and as plam points out, those too are worked to prevent, by installing guard rails, nets, phones, etc.  On our roads, it isn't, it's an accepted part of the design.

If we accept death, then we will have deaths, if we choose not to, we can work to eliminate them.  The Swedish have done it, they have done it without banning cars everywhere (although some places, where it makes sense), they have done it in a sparse country with a lot of rural roads and driving.  To pretend that we cannot is a lie, and it is only to justify killing people out of fear of being inconvenienced.

[MidTowner]

Can I humbly suggest that this might be too specific a discussion now for the General Road and Highway thread? Maybe we should have a thread about highway safety, or the proposed increases to speed limits on the 400-series.

Also, not to whinge, but while I am, I think the Grand River Transit thread merits designation as an 'Important Thread' as much as this one does.

[danbrotherston]

Can I humbly suggest that this might be too specific a discussion now for the General Road and Highway thread? Maybe we should have a thread about highway safety, or the proposed increases to speed limits on the 400-series.

Also, not to whinge, but while I am, I think the Grand River Transit thread merits designation as an 'Important Thread' as much as this one does.

Yeah, this is a good point.  I'm a bit surprised there is no vision zero thread already.

[robdrimmie]

I have little to contribute to this specific discussion, but I have found it very enlightening even though it has been difficult. Thank you to the folks who have worked through challenging differences of opinion.

[skyrise32]

And, of course, this idea that road fatalities are just accepted and never investigated is also ridiculous.  Just because we don’t have the resources to investigate every accident in detail like we do for major plane accidents doesn’t mean it never happens.

Major accidents (like say the Humboldt Broncos accident) are looked at in depth and recommendations are made to try to prevent similar accidents from ever happening again.  Wide scale accident statistics are looked at and acted on.  For example, we have concrete barriers for most of our divided highways because we found a common cause of death was crossing over through the ditch into oncoming traffic.

I did not say "never", I said, when the number does not exceed the expected number.  It's literally exactly what the city does.  The region has a list of intersections that they investigate for collisions.  That list is ordered by how much the injury rate EXCEEDS THE EXPECTED NUMBER.  I.e., how many more people are killed beyond the number they expect to be killed--you know, the number that it's okay to kill or injure.

Go ask a traffic engineer, they have actual numbers for the expected number of injuries and deaths for different intersections in the city, including brand new ones.  This is how the field works.

The situations you describe are when there was "more than expected" collisions.  I am not joking here, I'm not exaggerating, I am being absolutely serious. Traffic engineers have an rate of injury and death that is considered normal, and only in cases where that rate is exceeded is there a "problem".

Tou think that all engineering disciplines are the same, that isn't true, no bridge designer has an acceptable number of collapses greater than zero.  Yes, sometimes bridge do collapse, due to mistakes, or lack of repair, or misunderstanding the design, but those are considered failures.  Yes, sometimes there are suicides, and as plam points out, those too are worked to prevent, by installing guard rails, nets, phones, etc.  On our roads, it isn't, it's an accepted part of the design.

If we accept death, then we will have deaths, if we choose not to, we can work to eliminate them.  The Swedish have done it, they have done it without banning cars everywhere (although some places, where it makes sense), they have done it in a sparse country with a lot of rural roads and driving.  To pretend that we cannot is a lie, and it is only to justify killing people out of fear of being inconvenienced.

You are comparing apples to oranges with bridge design safety factor to what you are saying in regards to roads. Obviously no bridge designer has an acceptable number of collapses. In the same light, roads are not designed to fail either...they are built to strict standards and inspections and don't collapse (except for outside factors such as watermain leaks causing sinkholes), in regards to highways you can trust that the road will not fail and the barriers will take a hit from a car.
Its the human factor that traffic engineers have to take into account, one person will fall asleep at the wheel, bad drivers, speeding, drunk driving... which will always happen out of their control. Same people will use the bridge and possibly kill a pedestrian in the same way.
A bridge engineer and road engineer both design the same way..to take the expected load and last x number of years, its the municipality/mto that will decide if its safe, additional safety factors, speed limits etc.

[SammyOES]

I'm not going to bother touching the "expected number of fatalities" thing above.   It's been covered.  There are many human activities/products where absolute safety is so impossible that there is a background expected probability of failure/injury.  So, once again, this isn't special for road design.

Tou think that all engineering disciplines are the same, that isn't true, no bridge designer has an acceptable number of collapses greater than zero.  Yes, sometimes bridge do collapse, due to mistakes, or lack of repair, or misunderstanding the design, but those are considered failures.  Yes, sometimes there are suicides, and as plam points out, those too are worked to prevent, by installing guard rails, nets, phones, etc.  On our roads, it isn't, it's an accepted part of the design.

Um... the "acceptable" number of suicides isn't zero.  And if you think every suicide is investigated and acted on, you're just wrong.  It's exactly the same as road deaths, some number of deaths is treated as expected.  It's only when rates rise above the background rate that action is even considered.  

If we accept death, then we will have deaths, if we choose not to, we can work to eliminate them.  The Swedish have done it, they have done it without banning cars everywhere (although some places, where it makes sense), they have done it in a sparse country with a lot of rural roads and driving.  To pretend that we cannot is a lie, and it is only to justify killing people out of fear of being inconvenienced.

We can work to eliminate deaths even as we accept the reality that there will always be deaths.  I don't need to pretend that moving massive number of humans, large distances, at crazy high speeds (relative to what our bodies are designed for) is something that we can ever accomplish without accidents.  I can still say that we should work to improve safety.

Not to mention, that that kind of delusion, could easily be BAD for overall safety.  Ranking intersections by injuries over expected injuries seems like a good way to find problematic places to focus on new safety initiatives.

[danbrotherston]

I'm not going to bother touching the "expected number of fatalities" thing above.   It's been covered.  There are many human activities/products where absolute safety is so impossible that there is a background expected probability of failure/injury.  So, once again, this isn't special for road design.

Tou think that all engineering disciplines are the same, that isn't true, no bridge designer has an acceptable number of collapses greater than zero.  Yes, sometimes bridge do collapse, due to mistakes, or lack of repair, or misunderstanding the design, but those are considered failures.  Yes, sometimes there are suicides, and as plam points out, those too are worked to prevent, by installing guard rails, nets, phones, etc.  On our roads, it isn't, it's an accepted part of the design.

Um... the "acceptable" number of suicides isn't zero.  And if you think every suicide is investigated and acted on, you're just wrong.  It's exactly the same as road deaths, some number of deaths is treated as expected.  It's only when rates rise above the background rate that action is even considered.  

If we accept death, then we will have deaths, if we choose not to, we can work to eliminate them.  The Swedish have done it, they have done it without banning cars everywhere (although some places, where it makes sense), they have done it in a sparse country with a lot of rural roads and driving.  To pretend that we cannot is a lie, and it is only to justify killing people out of fear of being inconvenienced.

We can work to eliminate deaths even as we accept the reality that there will always be deaths.  I don't need to pretend that moving massive number of humans, large distances, at crazy high speeds (relative to what our bodies are designed for) is something that we can ever accomplish without accidents.  I can still say that we should work to improve safety.

Not to mention, that that kind of delusion, could easily be BAD for overall safety.  Ranking intersections by injuries over expected injuries seems like a good way to find problematic places to focus on new safety initiatives.

Why would this be better than ranking them by absolute number of deaths and injuries?  What is the goal?  Achieve no deaths or injuries, as Sweden aims to do?  Because if that's the goal then that is not a good way to rank intersections.  In fact, that ranking seeks to achieve the goal of kill or injure this number of people.

Please, answer for me, why are people in Sweden not willing to kill people, but we are?

Maybe I should move there...

[tomh009]

Please, answer for me, why are people in Sweden not willing to kill people, but we are?

Sweden set the Vision Zero goal (zero traffic deaths by 2020) in 1997, 22 years ago. They have reduced traffic deaths by about 50%, but the fatality rate has now been pretty much flat for the last five years.

So the Swedes are still willing to "kill" 250 people per year.

Your choice of the verb "to kill" is (probably intentionally) highly provocative/incendiary, given that it means "to cause the death of a person". And I think most people would agree that road design is not the cause of all (or even most) traffic deaths in Canada, or in other countries.

[danbrotherston]

Please, answer for me, why are people in Sweden not willing to kill people, but we are?

Sweden set the Vision Zero goal (zero traffic deaths by 2020) in 1997, 22 years ago. They have reduced traffic deaths by about 50%, but the fatality rate has now been pretty much flat for the last five years.

So the Swedes are still willing to "kill" 250 people per year.

Your choice of the verb "to kill" is (probably intentionally) highly provocative/incendiary, given that it means "to cause the death of a person". And I think most people would agree that road design is not the cause of all (or even most) traffic deaths in Canada, or in other countries.

No, the Swedes *ARE* killing 250 people a year, which they consider unacceptable, and believe that the number should be zero.

We are killing ~2000 people per year, and with respect to road design, consider that number perfectly acceptable...or more specifically, are not willing to make any meaningful changes to road design which could reduce that number.  We sure are willing to blame drivers, pedestrians, cyclists, weather even, but the absolute refusal to make safety the priority when designing roads is the single defining difference between Sweden and here.

And yes, Sweden has not completed their Vision Zero project.  I doubt they ever will, just as I doubt there will be a day when we never have aviation disasters.  That does not mean it should not be the goal.  Further, if our goal had been to get deaths to zero 20 years ago as well, does that mean we could be killing merely ~1500 people now? Which of those 500 people who would be alive per year today are worth the sacrifice?

And yes, I am using incendiary language because this is an incendiary issue. It is unacceptable to me that people should die on our roads, and it is inconceivable to me that others think this is okay.

As for the cause, I don't care about the cause, nobody should care about the cause, they should care about what can be done to mitigate and reduce the risk.

Please consider a specific example, which I really like.  One of the most serious types of collisions is crossover collisions on rural roads, they're high speed which leads to high fatalities.  In Sweden most rural two lane highways built or reconstructed in the past 20 years have been rebuilt with a cable barrier in the middle. This restricts passing to sections which have passing zones.  This means that if a driver is doing 75, we all must go 75.  It's inconvenient, annoying, and when there are farming vehicles, probably a cause of a few legitimate delays.  It does increase the cost of the road, cables aren't free, and building more passing zones isn't so cheap (although Sweden usually just used the existing road shoulder).  Yet it's been incredibly effective in nearly eliminating crossover collisions as they are no longer possible.  How do people here respond to that idea?  Both, in this forum, and in the region in general?  I don't care that the cause is aggressive angry driving. I don't think we can ever fix that, although, if you have a way, I'm happy to hear it.  But these barriers would help mitigate those collisions at reasonable cost, and reasonable inconvenience.  Why would we oppose them.

[tomh009]

And yes, I am using incendiary language because this is an incendiary issue.

OK. Please enjoy the incendiary discussion.

[neonjoe]

I think the number of expected injuries is plainly due to the fact that the squishy things between the wheel and the seat bring upon too much uncertainty to allow for a zero rate. Those squishy things often operate the vehicle outside of the norms for the designed conditions, speed, acceleration, deceleration, attention. Self driving cars will make the roads safer but there’s always the human element involved. A bridge isn’t designed to fail but if human hijacks a tank and runs into it it may, if the engineer had this case as part of the variables included in the tolerances we would either have designs that are much much heavier or there would be a tolerance for failure based on the fact that tank collisions are now part of the design variables.

[SammyOES]

Why would this be better than ranking them by absolute number of deaths and injuries?  What is the goal?  Achieve no deaths or injuries, as Sweden aims to do?  Because if that's the goal then that is not a good way to rank intersections.  In fact, that ranking seeks to achieve the goal of kill or injure this number of people.

Please, answer for me, why are people in Sweden not willing to kill people, but we are?

Maybe I should move there...

It’s not an exclusive choice we have to make.  We can (and do!) do both.  But if you have a low usage intersection that is producing a relatively large number of fatalities there - we should look at that!  There’s more likely to be low hanging fruit changes that can be made there.  

Remember, our resources are limited and perfection is impossible.

People in Sweden are willing to kill people. That why they still have a lot of fatalities (in an absolute sense) from roads and are in no way close to actually having zero fatalities.

You can’t give the majority of a population complete control over a high mass device that can travel at speeds of 100km/h or more and actually think there’s a way to never have fatalities.  Sorry.

Edit: Reading the other posts it seems pretty clear this isn’t worth continuing. I think peoples actions reflect their beliefs and not the words they use. Semantics means little to me if it doesn’t change actions. Sweden can *say* they don’t accept casualties but their actual actions don’t reflect that.

As for the barrier idea, it’s probablu a good idea in many places here. It’s probably a bad idea in many other places. It again comes down to a cost/benefit analysis.

[danbrotherston]

I think the number of expected injuries is plainly due to the fact that the squishy things between the wheel and the seat bring upon too much uncertainty to allow for a zero rate. Those squishy things often operate the vehicle outside of the norms for the designed conditions, speed, acceleration, deceleration, attention. Self driving cars will make the roads safer but there’s always the human element involved. A bridge isn’t designed to fail but if human hijacks a tank and runs into it it may, if the engineer had this case as part of the variables included in the tolerances we would either have designs that are much much heavier or there would be a tolerance for failure based on the fact that tank collisions are now part of the design variables.

Tell this to the Swedes.  They'll tell you that designing for human factors is entirely possible.