2013 saw the third lowest number of deaths on Northern Ireland’s roads since records began. The provisional figure of 56 deaths is higher than last year – which saw 48, the lowest ever – but is still in keeping with the trend that has been set in the past few years. In this blog I unpick the statistics to see what trends we can identify.
I should stress at the outset that these are not just statistics. Every person who dies on the road is someone’s child and loved one, so nothing in this blog is intended to in any way trivialise the lifelong grief that has resulted from every single one of these deaths.
Fears Earlier in the Year
During the Spring a number of concerns were expressed (including by myself) about the rate of deaths in 2013. By the end of April the number of deaths up to that point in 2013 (17) was more than double the rate in the first four months of 2012 (8). However the final figure for the year was not anything like as different, and we can now see why. Roads deaths in 2012 were focused towards the latter part of the year, whereas in 2013 they were more evenly distributed through the year:
So we can now see that, in fact, the concern was unwarranted since the anomaly was not the high number of deaths in the first four months of 2013, but rather the low number of deaths in 2012. This is a good illustration of why we should be careful not to read too much in to short term trends. Road deaths are not ‘random’ events in the sense that most could be avoided. However, in mathematical/statistical terms they are ‘random’ in that the occurrence of one road death on a given day has no bearing on the occurrence of another road death on another day. Therefore, we should expect to see clusters of deaths in given months but this should not cause us alarm unless the trend seems to be continuing beyond the month or two in question.
Therefore, the identification of trends and conclusions is more robust over longer time periods. So therefore, in this analysis I have decided to combine the figures for 2012 and 2013 and analyse them together (except, obviously, for the first section where I look at total deaths per year). This ought to reduce the number of false, i.e. short term, trends and more accurately highlight longer term trends.
Deaths By Year
Headline: The recent decline in roads deaths seem to have levelled off.
The graph below shows the number of deaths per year on Northern Ireland roads over the past 17 years. From this, we can see that road deaths stood steady at approximately 150 per year from 1996 to around 2004. The period 2004 to 2010 saw an unprecedented decline from 150 per year down to around 55 per year, something that we can be justifiably pleased about, although not complacent. It now seems that since 2010 the decline has halted and the death rate per year is now holding steady once again, this time at an average of 55 per year.
In my 2012 analysis I offered some more detailed discussion about why this remarkable decline occurred between 2004 and 2010, but it appears to have been a combination of (a) better car design (b) road improvement schemes (c) anti-speeding campaigns and (d) anti-drink/drug driving campaigns. Three years ago the Department of the Environment published the Northern Ireland Road Safety Strategy to 2020. This contained an ambitious target to reduce road deaths by at least 60% compared to the 2004-2008 average by the year 2020. This would mean reducing the average road deaths per year to 50. So with a four-year average of 55 we are very close to achieving this. More can still be done. Nor should we stop the campaigns which contributed to the recent decline, since we need to hold on to these gains.
Significance of Deaths Being Higher than Last Year
The “easy” headline to write would be “road deaths up on last year”. This would be factually correct, but I don’t believe there is sufficient justification to highlight this as the key point to be made about road deaths in 2013. As explained above, road fatalities are statistically ‘random’ so will always tend to fluctuate from year to year, sometimes rising and sometimes falling. The long-term trend is what we need to watch. The 56 road deaths in 2013 was only slightly higher than the four-year average of 55, so I see no reason to be unduly concerned that we have seen more than this over the past twelve months. The most likely explanation is that 2012 was unusually low, and therefore the rise we have seen this year is probably just an illustration of regression toward the mean and not something we ought to put an unwarranted focus on. It would only start to be concerning if the next 2 or 3 years saw a continued rise away from the 2012 low point rather than remaining close to the average.
Deaths By Gender
Headline: Males are almost three times as likely to die on our roads than females.
The Northern Ireland Travel Survey shows that males travel almost 30% more than females. But even taking this into account, males were killed at a rate of 4.24 for every billion kilometres they travelled, compared to 1.67 for females. This suggests that men are inherently more likely to be killed on the road than women. There could be a number of reasons for this such as: different driving styles between men and women; or perhaps the fact that a higher proportion of men use types of vehicle that have poorer safety records in Northern Ireland, such as motorcycles or bicycles. This suggests that in our efforts to reduce deaths further, measures aimed particularly at men would seem sensible.
Deaths By Age and Gender
Headline: Men under 30 and over 60 are the most likely to die on our roads.
The statistics show a significant difference between different age/gender groups. The graph below shows the death rate per billion kilometres travelled for four age groups – children under 16, 16-29, 30-59 and 60+ – and also by gender. Stating these in terms of deaths per billion kilometres travelled, rather than the total number of deaths, cancels out the fact that the old and young tend to travel less than those of middle age.
These figures show two strong spikes in deaths, amongst young males aged 16-29 and men over the age of 60, both of whom experience more than twice the average death rate of 3.1 deaths per billion km travelled. By contrast the death rate among women in these two age groups is around average. Middle aged people (30-59) have the lowest death rates amongst adults, while the death rate amongst children is, thankfully, well below average. The safest group of road users is women aged 30-59, whose death rate is significantly below average. This all suggests that when making further efforts to reduce road deaths, measures aimed at people aged between 16-29 and over the age of 60 would seem prudent, and in particular males in these age brackets.
The reasons for the high deaths rate in the two age groups highlighted above are likely to be different. Amongst men aged 16-29, inexperience and a tendency to riskier behaviour on the road is likely to play a more significant role than amongst older men, while for men aged over 60 the frailty that comes with advancing years is likely to be a more significant factor than it would be for younger men. However I don’t have the data to be able to say more than this on the matter.
Deaths By Mode of Transport
Headline: Motorcyclists have by far the highest fatality rate, followed by cyclists and pedestrians.
The pie chart below shows the proportion of deaths that occurred for each category of road user during 2012 and 2013. As you can see, car occupants make up the highest proportion at 57%, while pedestrians and motorcyclists represent the next highest at 15% and 13% respectively.
However, what this masks is the significantly different proportions that these different modes of transport make up on actual roads. Cars, for example, are the dominant means of transport in Northern Ireland, representing far more than the 57% of journeys that the death statistics would imply. Motorcyclists, by contrast, only represent a minuscule proportion of traffic, about 0.1% of all distance travelled in the province. They are thus significantly over-represented in the fatalities. Similarly, although cyclists “only” made up 5% of fatalities, they make up only 0.5% of distance travelled in Northern Ireland, so they too are over-represented in the fatalities.
A better approach is to calculate the number of deaths per billion kilometres travelled for each mode of transport in Northern Ireland. This gives us a measure of how likely a given road user is to be killed while using that mode of transport. The results are quite shocking. Remember, the average death rate on Northern Ireland roads in this time period was 3.1 deaths per billion kilometres travelled.
What these figures show is that a motorcyclist is almost 100 times more likely to be killed on our roads than the average person. Cyclists are around ten times more likely to be killed, and pedestrians six times more likely to be killed. Occupants of HGVs and other vehicles (such as tractors and vans) are just slightly higher than average, while a car occupant is somewhat less likely to be killed than the average person, making cars the second safest means of travel on the road, despite more than half of all deaths being of car occupants. Buses remain our safest mode of road transport, there having been no deaths amongst bus passengers in the past two years.
In some ways, these figures are not surprising. What they show is that road users with lots of metal around them (buses, lorries and cars) are far less likely to be fatally injured in a crash than a road user with very little (cyclists and motorcyclists) or none (pedestrians). Amongst those with little or no metal around them, the road users who are capable of the highest speeds (motorcyclists) have the highest death rate. This all suggests that a key element of the survivability of an accident is the amount of other “stuff” sitting between the person and whatever is impacting them, and for those with not much “stuff”, the speed of the impact is critical. This is really just common sense.
So further efforts need to be made to increase the safety record amongst motorcyclists, cyclists and pedestrians in particular. The way that this can be achieved will differ greatly for each of the three road users. Motorcycles tend to have high cruising speeds, which certainly adds to the risk of a fatality in the event of a collision. At the same time, there is a history of motorcyclists being struck by car drivers who were changing lanes or emerging from side roads and who did not see the motorcyclist because they were not paying adequate attention. So for motorcycles, the death rate can be reduced by continuing to encourage a ‘bike safe’ culture and by increasing driver education.
For cyclists, attention will focus on how safe the road network is for those on bicycles, especially in somewhere like Northern Ireland where cars are so dominant that drivers often do not expect to encounter a cyclist. A second focus will be on the interaction between larger vehicles and cyclists. In 2012 and 2013, 2 cyclists were killed by collision with cars, 1 with a van, 1 with a bus and 1 with an ambulance. Larger vehicles therefore represent a higher proportion of cyclist fatalities than would be expected from the number and type of vehicles on the road. Unlike pedestrians, who largely have their own segregated infrastructure in the form of footpaths, cyclists generally have to share space with motor vehicles. So cyclist deaths could be tackled through a combination of driver education and improved – ideally segregated – cycling infrastructure.
Unlike other road users, the majority of deaths among pedestrians occur in urban or suburban settings with 30mph speed limits. Children are sadly over-represented in pedestrian deaths in Northern Ireland (making up 22% of all pedestrian deaths, compared to 7% across all modes of travel). So further attempts to reduce pedestrian deaths should focus on urban areas, and particularly on children.
All of this notwithstanding, the majority of deaths still occur among car occupants. So while the safety record of cars is already quite good, there is still plenty of scope for further reductions. The majority of car crashes occur due to carelessness: things like not looking where you are going, being in the wrong lane or not looking carefully enough when turning in or out of side roads. This can be tackled partly through physical road improvements, such as sightline improvements and dual-carriageway upgrades and partly through continued efforts to increase driver education. Road safety adverts over the past two years have correctly identified this factor and focused on inattention as the number one killer on the roads.
Comment added 1 Jan 2014: After it was pointed out that “deaths per billion km travelled” is only one of the ways to measure fatalities, I have added a new section at the end of this blog giving two alternatives.
Deaths By Setting
Headline: Most road deaths occur in rural areas.
Despite the majority of journeys in Northern Ireland happening in urban and suburban settings, the vast majority of road deaths occur in rural areas as shown in this chart. (“Urban” means a setting where buildings are typically located right at the edge of the road, such as in a town centre. “Suburban” means a setting where buildings are typically separated from the edge of the road by a garden or other green space).
Similarly, the majority of road deaths occur on roads with a maximum 60mph speed limit:
In some ways this is not surprising, given that that vast majority of the Northern Ireland road network has a maximum 60mph speed limit. But 60mph speed limits tend to be located in rural areas, again highlighting the tendency of deaths to be concentrated in rural areas. Although it is important to make ongoing efforts to reduce deaths in urban areas, the fact that the majority still occur in rural areas suggests that the latter that also needs to have a lot of attention moving forwards. So we need to focus in reducing deaths on rural roads.
The fact that vehicles can attain higher speeds on rural roads is surely a factor. A vehicle leaving the road and striking a tree at 60mph is much more likely to result in a fatality than a similar impact at 30mph. There is also a tendency amongst some road users to travel at inappropriate speeds on rural roads, particularly unclassified roads (i.e. those minor roads that are not M, A or B class).
A blanket maximum 60mph speed limit applies by default to all single-carriageways roads without a specific lower limit. Rural roads are often twisty and undulating, and corners can bring sudden surprises such as tractors, livestock, pedestrians or mud. I would suggest that 60mph is not a safe speed to travel on the majority of the rural road network in Northern Ireland. The 60mph limit merely means that this is the legal maximum speed, not that 60mph is a safe speed to travel on these roads. There is a strong argument that the national speed limit should be only be set at 60mph on A-class roads, but that all other rural roads should have a speed limit of 50mph, similar to the situation in the Republic of Ireland where rural roads have an 80 km/h limit. However, since it is highly unlikely that the PSNI would have the resources to enforce a 50mph speed limit in even the most basic manner across the vastness of the rural road network (research shows that 40% of drivers in RoI flout the 50 limit), driver education is still the avenue that is likely to lead to the most benefits, so we should focus on further driver education on driving on rural roads.
By Road Standard
Headline: Motorways and dual-carriageways remain our safest roads.
The graph below shows the death rate per billion kilometres travelled for each of four categories of road in Northern Ireland. (Unlike the previous graphs, this one is based on vehicle mileage, not the mileage of all road users. So passengers in cars, for example, are not counted. This is why the death rates are all below the average of 3.1 for all modes of transport. However the relative differences between each type of road would be unchanged if this was included, and this is the key point to the graph.)
Motorways, despite having the highest average speeds, are still our safest roads with A-class dual-carriageways not far behind. This demonstrates that it is not simply speed that kills, but rather speed in an inappropriate setting. High speed driving is appropriate on these high-quality roads since the road is designed with this in mind and has been engineered to greatly reduce the possibility of accidents. Motorways were designed to allow safe travel at sustained high speeds, and the statistics show that they are still achieving this aim. Higher fatality rates occur on single-carriageway roads, which tend have more opportunities for crashes to occur, and more severe consequences if they do (e.g. head-on crashes). A-class single-carriageways are more dangerous than B-class roads and below, perhaps due to the fact that these roads tend to be busier and have higher average speeds, and which present more opportunities for drivers to be come over-stimulated and miss crucial pieces of information.
The following basic points can be made about road deaths over the past two years:
- 56 people died on Northern Ireland’s roads in 2013, the third lowest on record.
- Road deaths seem to have levelled off at an average of 55 deaths per year.
- Males are almost three times as likely to die on the roads as females.
- Men aged 16-29 and men aged over 60 are the most likely age/gender group to die on the roads.
- Motorcyclists are almost 100 times more likely to die than the average road user.
- Cyclists and pedestrians also have a significantly higher death rate than average.
- Children are over-represented in pedestrian fatalities.
- Around three-quarters of all deaths occur in rural areas.
- Motorways and dual-carriageways are still our safest roads.
Wishing everyone a safe and enjoyable 2014 on the roads.
Added 1 Jan 2014: Other Measures of Fatalities Per Mode
After I published this blog it was commented by one reader out that deaths per billion kilometres travelled is only one way of measuring fatalities. Two others are (b) fatalities by number of journeys made and (c) fatalities by hours travelled. All measures have pros and cons, but for now suffice it to say that each measure has its pros and cons. The one you choose depends on whether you think of danger in terms of (a) the risk of being killed while travelling a certain distance (b) the risk of being killed while undertaking a given journey or (c) or the likelihood of being killed in a given hour of travel. So for this reason I am showing graphs for Northern Ireland fatalities by all three measures below. The first graph is the same as the one above, reproduced here for convenience.
The average fatality rates across the whole population are as follows: 3.1 deaths per billion km travelled, 32 deaths per billion journeys made and 99 deaths per billion hours travelled. The results give slightly different indications of how ‘safe’ each mode of transport is when measured in terms of fatalities.
- Car users. There is not a huge difference between the three measures, all of which show cars to be slightly safer than average. ‘Deaths per billion km travelled’ gives the ‘safest’ measure at 67% of average, while the ‘deaths per billion hours travelled’ gives a figure that is 81% of average.
- HGV/Other Vehicles. Slightly more variation here. ‘Deaths per billion km travelled’ gives the ‘safest’ measure at about 1.3 times higher than the of average, while the ‘deaths per billion journeys’ suggests the risk is about 2.3 times higher than average.
- Pedestrians. The ‘Deaths per billion km travelled’ ranks pedestrians as 6 times more likely to die than average. However the other two measures suggest that the risk is much lower, around average for ‘deaths per billion journeys’ and slightly below average for ‘deaths per billion hours travelled’.
- Cyclists. The ‘Deaths per billion km travelled’ ranks cyclists as 10 times more likely to die than average. However, as with pedestrians, the other two measures suggest that the risk is lower, at around 6 times the average for ‘deaths per billion journeys’ and 5 times the average for ‘deaths per billion hours travelled’.
- Motorcyclists. They still come out as the most at-risk group of road users with each analysis, but the level of risk varies. The ‘Deaths per billion km travelled’ ranks motorcyclists as 100 times more likely to die than average. The ‘deaths per billion journeys’ ranks it even higher at 120 times the average, while the ‘deaths per billion hours travelled’ ranks it a bit lower at 39 times the average.
Sources of data:
- For details of road deaths: my own records of road deaths compiled from media reports, verified against the PSNI’s daily fatal statistics, updated daily.
- For distance travelled by different types of user: Travel Survey for Northern Ireland In-depth Report 2010-2012, published December 2013.
- For population of Northern Ireland and gender/age breakdown: 2011 Census, NISRA
- For kilometres travelled by different classes of vehicle: 2009 Roads Service Traffic and Travel Information Report (the most recent available).
- For kilometres travelled on different classes of road: 2009 Roads Service Traffic and Travel Information Report (the most recent available) but supplemented by my own calculations to separate out dual A-class roads and single-carriageway A-class roads.