the licensing and safety of older drivers in britain
TRANSCRIPT
T
C1
a
ARRA
KDCSRF
1
ac1o
pdvwda
mfiAdiio
0h
Accident Analysis and Prevention 50 (2013) 732– 741
Contents lists available at SciVerse ScienceDirect
Accident Analysis and Prevention
jo ur n al hom ep a ge: www.elsev ier .com/ locate /aap
he licensing and safety of older drivers in Britain
hristopher G.B. (Kit) Mitchell ∗
7 Tavistock Road, Fleet, Hampshire GU51 4EH, UK
r t i c l e i n f o
rticle history:eceived 27 March 2012eceived in revised form 20 June 2012ccepted 27 June 2012
eywords:river licence forecastar driver fatality forecastafety older drivers
a b s t r a c t
Driving licence holding in Britain is increasing for older people, particularly older women. Licence holdingby men aged 40–69 has saturated at about 90%, and for women aged 30–59 at about 78%. Drivers beginto surrender licences after age 70. By age 90, 38% of women and 25% of men who held a licence at age 70have surrendered their licences.
The paper shows the numbers of car drivers of different ages killed and injured in accidents since 1975and the fatality and casualty rates per driver and per mile driven. The safety of older drivers is improvingfaster than that of younger age groups. The number of car driver fatalities aged 70 and over was highestbetween 1990 and 2004, and has subsequently reduced by almost 40%. For drivers aged 80 and over, the
isk to other road usersragility of older persons
peak was in 2004 and the number has subsequently reduced by almost 50%.The paper uses demographic projections, forecast licence holding and the trends in fatality rates to
project the numbers of fatalities for drivers of different ages in Britain. This shows that fatalities amongolder drivers aged are likely to continue to reduce.
Fatality rates for older road users are increased by the fragility of older persons. This disguises the factthat the accident involvement rate for older drivers does not increase with age until after age 75 or 80.
. Introduction
The increase in the population of older people is well known,nd is occurring in Britain in a way similar to virtually every otherountry. Britain currently has an older population than USA, with6.6% of the population aged 65 and over and 2.3% aged 85 andver, compared with 13.0 and 2.0% respectively in USA.
Not only has the older population been increasing, but theercentage of older people, particularly older women, who holdriving licences has also been increasing. Older people are moreulnerable to accidental injury because physical fragility increasesith age (Evans, 2000). In the 1980s and 1990s researchers pre-icted a large increase in the number of older drivers killed in trafficccidents (TRB, 1988; Burkhardt and McGavock, 1999).
In practice, in both USA and Britain, since these predictions wereade the number of elderly driver fatalities has peaked and is now
alling, because the rate of casualties per driver for older car driverss reducing faster than their number is increasing (see Section 3).t the same time, research is showing that although the oldestrivers have a high fatality rate per mile driven, their rate for slight
njuries per mile driven, which is a proxy for accident involvement,ncreases much less. Most of the increase in the fatality rate forlder drivers is a result of their fragility increasing with age, rather
∗ Tel.: +44 1252617621.E-mail address: [email protected]
001-4575/$ – see front matter © 2012 Published by Elsevier Ltd.ttp://dx.doi.org/10.1016/j.aap.2012.06.027
© 2012 Published by Elsevier Ltd.
than their accident involvement increasing. Also, because driversreduce the mileage they drive as they age, their accident involve-ment rate per year does not increase with age below 75 or 80 years,and rather little for ages beyond 80 years. In addition, older cardrivers are shown to represent a small danger to other road users,in that they kill a small percentage of all the pedestrians killed bycars. This is leading to the view that while older drivers are at riskthemselves, because of their fragility, they are no greater threat toother road users than are younger drivers (Hakamies-Blomqvist,2004).
There has been extensive research on older drivers since the TRBreport on transportation in an ageing society (TRB, 1988). A recentsynthesis of research has been provided by Eby et al. (2008). Thiscontains sections on medical conditions and medications, screen-ing and assessment, education and rehabilitation, vehicles andadvanced technology, roadway design and transitioning to non-driving. An overview of the situation in Britain is provided by Boxand Mitchell (2010).
Eustace and Wei (2010) examine motor vehicle fatal crashesin detail to assess the contribution of age and gender to drivingerrors that cause fatal motor vehicle crashes, using data from FARS(NHTSA, annual) for the three-year period 2001–2003. This showsthat the proportion of crashes for which the driver was held to be
responsible rises steadily after age 50, reaching 80% by age 85 andover.Data on the proportion of drivers at fault by age are not generallyavailable in Britain, so the type of analysis made by Eustace and Wei
lysis and Prevention 50 (2013) 732– 741 733
(tlddc
irnsddie
2
2
Ndfs
dal3Tlb
hchlsf
1t1wf
a
2
twftt
alvu
qs
0
10
20
30
40
50
60
70
80
90
100a
1965 197 0 197 5 198 0 198 5 1990 1995 2000 2005 2010
Pe
rce
nt h
old
ing
fu
ll ca
r d
rivin
g lic
en
ce
s
.
17 - 20
21 - 29
30 - 39
50 - 59
60 - 69
70+
75 - 79
85+
Age range
0
10
20
30
40
50
60
70
80
90
100
1965 1970 1975 1980 1985 199 0 199 5 200 0 200 5 2010
Pe
rce
nt h
old
ing
fu
ll ca
r d
rivin
g lic
en
ce
s
.
17 - 20
21 - 29
30 - 39
50 - 59
60 - 69
70+
75 - 79
85+
Age rangeb
Fig. 1. (a) Percentage of men holding car driving licences – Great Britain National
C.G.B. (Kit) Mitchell / Accident Ana
2010) is not possible using national statistics. This paper addresseshe issue of whether the increase in older drivers in Britain willead to an increase in older car driver casualties and whether olderrivers pose a larger danger for other road users than do youngerrivers. To do this, it predicts trends in licence holding and car driverasualties in Britain to 2030.
It shows the growth of car driving licences among older driversn Britain, with evidence on the rate at which older drivers sur-ender licences as they age. It uses this information to predict theumber of older drivers up to year 2030. It then demonstrates howafety has been improving for car drivers of various ages and pre-icts the number of fatal casualties to 2030. The number of olderriver fatalities is likely to continue to fall. To put the numbers
n context, older driver fatalities will continue to be substantiallyxceeded by fatalities among young drivers.
. Car driving licence holding
.1. Trends in licence holding by men and women
Data on car driving licence holding in Britain come from theational Travel Survey, a household survey that has been con-ucted regularly since 1972 and annually since 2002 (Departmentor Transport, annual (a)). At present about 20,000 individuals areurveyed and complete a one-week travel diary.
From these surveys it is possible to follow cohorts of respon-ents (but not individuals) and deduce when people obtain licencesnd also how many surrender licences as they age. Published tabu-ations provide data on licence holding by age groups 17–20, 21–29,0–39, 40–49, 50–59, 60–69 and 70 and over. The Department forransport has supplied the author with special tabulations givingicence holding for five-year age groups from 50–55 to 85 and over,ut only for years since 1995–1997.
Fig. 1a and b shows the percentages of men and women whoeld car driving licences since 1965 (age group 40–49 omitted forlarity). It can be seen that for men aged up to 70, licence holdingas stabilised at around 90%. For older age groups, licence holding is
ower but increasing. For women aged up to 60, licence holding hastabilised at just below 80%. Again, holding is lower but increasingor older age groups.
For both men and women aged 17–20, licence holding peaked in992–1993, fell back rapidly until 2004 and is now increasing again,hough it appears to be stabilizing at levels lower than the peak in992–1993. This has now reduced the licence holding of men andomen aged 20–29. It is beginning to reduce it for men aged 30–39,
or whom licence holding has fallen from 90% in 1992–1993 to 85%.The large increase in licence holding by both men and women
ged 70 and over since 1975 is clear.
.2. Licence renewal procedure
Having passed a test of driving competence, which includes aheory section and a practical test on public roads, a driver is issuedith a car driving licence valid to age 70. The driver is responsible
or reporting to the Driver and Vehicle Licensing Agency (DVLA) ifhey experience a medical condition or injury that would preventhem from holding a driving licence.
At age 70, on completing a self-declaration of not suffering from medical condition that would prevent the driver from holding aicence, the car driver is issued at no cost with a renewed licence,alid for three years. This process is continued every three years
ntil the driver chooses not to renew their licence.The health declaration consists of answering about two dozenuestions by ticking boxes. Each question asks whether the driveruffers from a specific condition that would either represent a bar to
Travel Survey (Department for Transport, annual (a)). (b) Percentage of womenholding car driving licences – Great Britain. National Travel Survey (Departmentfor Transport, annual (a)).
holding a licence or alert DVLA to the need for a medical examina-tion. Sweden uses a very similar system for licence renewal, whileAustria, Belgium and France simply issue a licence for life. Stud-ies show that countries with relaxed procedures for the renewalof licences by older car drivers tend to have road systems thatare safer for older drivers. Hakamies-Blomqvist has shown this bycomparing Sweden and Finland (Hakamies-Blomqvist et al., 1996).Langford et al. (2004) has found similar results comparing the Aus-tralian states of Victoria and New South Wales, and Mitchell (2008)has compared a number of countries in the European Union.
2.3. Licence surrender in older age
By following cohorts it is possible to estimate how many driverssurrender or fail to renew their licences as they age. Tables 1a and 1bshow the percentage licence holding by for five-year age groups ofmen and women in 1985, 1990, 1996, 2000, 2005 and 2010 andthe changes in licence holding over various periods before vari-ous stated ages. These changes are for particular age cohorts. Thusin Table 1a, the change in licence holding over 10 years in 2005for 70–74 year olds is 84% (60–64 year olds in 1996) minus 79%(70–74 year olds in 2005). For 2010, the corresponding figures are85% in 2000 and 86% in 2010. Because the surveys follow cohorts butnot individuals, there is considerable random variation betweenyears.
The data for five-year age groups of older drivers were obtainedas special tabulations and are only available for 1996 onwards (data
for 1990 is from the 1989/1991 NTS, for 1996 from the 1995/1997NTS). From these data it is possible to estimate how licence holdingchanges as people age over 5, 10, 15 and 20-year periods. For ageingperiods that go back before 1996, the initial licence holding has
734 C.G.B. (Kit) Mitchell / Accident Analysis and Prevention 50 (2013) 732– 741
Table 1aChanges in licence holding with ageing – men, Great Britain.
Year Age
50–54 55–59 60–64 65–69 70–74 75–79 80–84 85+
Men – percentage holding licences1985 (10 year groups) 81 81 72 721990 (10 year groups) 85 85 78 781996 91 86 84 82 77 66 48 362000 87 88 85 86 78 68 51 442005 91 89 89 87 79 76 68 482010 87 92 90 89 86 81 73 55
Change over 5 years (for cohort with age at year stated)2000 −2 −3 −1 2 −4 −9 −15 −42005 2 1 2 −7 −2 0 −32010 1 1 0 −1 2 −3 −13
Change over 10 years1996 4 −1 3 1 5 −62000 −2 −1 −2 1 0 −102005 −2 1 −5 −6 −9 −182010 3 1 1 −5 −5 −13
Change over 15 years (approximate for younger age groups)2000 1 1 −2 1 −3 −4 −212005 2 0 0 0 −3 −2 −102010 −1 3 0 −2 0 −3 −9 −22
Change over 20 years (approximate for younger age groups)2005 5 3 2 0 −2 −5 −4 −242010 −1 4 1 1 1 −2 −5 −25
Age change Licence change Age change Licence change Age change Licence change Age change Licence change
Change in licence holding after ten years ageing – menAbsolute change % 63 → 73 −2 68 → 78 −5 72 → 82 −7 78 → 88 −15
−5
N
bhbo
TC
N
Percent of previous holding −2
ational Travel Survey (Department for Transport, annual (a)).
een estimated from data for ten year age groups. Where the author
as judged this to be insufficiently accurate, the table cell is leftlank. The increases in licence holding over longer periods is a resultf some people obtaining licences in middle age, before possiblyable 1bhanges in licence holding with ageing – women, Great Britain.
Year Ages50–54 55–59 60–64
Women – percentage holding licences1985 (10 year groups) 41 41 24
1990 (10 year groups) 49 49 33
1996 67 56 50
2000 71 64 58
2005 75 72 65
2010 79 76 73
Change over 5 years (for cohort with age at year stated)2000 4 −3 2
2005 1 1
2010 1 1
Change over 10 years1996 8 7 8
2000 5 −2 4
2005 −2
2010 2
Change over 15 years (approximate for younger age groups)2000 12 8 2
2005 9 6 −1
2010 5 3 0
Change over 20 years (approximate for younger age groups)2005 19 14 9
2010 12 18 7
Age change Licence change Age change Lice
Change in licence holding after ten years ageing – womenAbsolute change % 63 → 73 −4 68 → 78 −5
Percent of previous holding −7 −10
ational Travel Survey (Department for Transport, annual (a)).
−9 −20
surrendering their licence or allowing their licence to lapse later in
older age.At the bottom of the tables is shown the best estimate for theabsolute change in licence holding after ageing ten years by people
65–69 70–74 75–79 80–84 85+
243341 32 20 14 353 36 25 14 855 47 38 25 1364 53 45 32 19
3 −5 −7 −6 −6−3 −6 2 0 −1−1 −2 −2 −6 −6
−4 5 −64 3 −8
−1 −3 −3 −7 −70 −5 −8 −4 −6
5 −5 1 −10−3 −2 5 −7−3 −3 −5 −9 −13
7 6 6 1 −96 4 4 −1 −14
nce change Age change Licence change Age change Licence change
72 → 82 −6 78 → 88 −7 −17 −28
C.G.B. (Kit) Mitchell / Accident Analysis and Prevention 50 (2013) 732– 741 735
0
5
10
15
20
25
30
35
908580757065605550
Current age
Pe
rce
nt
Women - percent of licence holders who surrender in previous decade
Men - percent of licence holders who surrender in previous decade
Women - percent reduction in licence holding in previous decade
Men - percent reduction in licence holding in previous decade
Fc
ni
otismabwrihhl
2
iolhhbtw
7wo
ar
3
3
aGoio
i
0
10
20
30
40
50
60
70
80
90
100a
b
2030202020102000199019801970
Pe
rce
nt h
old
ing
lic
en
ce
s
17-20
21-29
30-39
40-49
50-59
60-69
70+
Age group
0
10
20
30
40
50
60
70
80
90
100
2030202020102000199019801970
Pe
rce
nt h
old
ing
lic
en
ce
s
17-20
21-29
30-39
40-49
50-59
60-69
70+
Age group
ig. 2. Reduction in licence holding over previous decade, and reduction as a per-entage of initial holding – men and women, Great Britain.
ow aged 73. 78, 83 and 88, and this change as a percentage of theirnitial licence holding at ages 63, 68, 73 and 78.
Fig. 2 shows the reduction in licence holding by men and womenf various ages in the 10 year period before the age plotted, andhese reductions as percentages of the initial level of licence hold-ng. There is no evidence of licence surrender before the age ofeventy, but after that age, a higher proportion of women thanen who hold licences surrender them. There are no data avail-
ble on the number of drivers who have their licences withdrawny the licensing authority as opposed to allowing them to lapsehen due for renewal, or surrender them voluntarily. Anecdotal
eports suggest that very few licences are withdrawn by the licens-ng authority. Because the data on licence holding come from aousehold survey, they are not affected by the deaths of licenceolders, as is the case for administrative records of the number of
icences.
.4. Forecast licence holding
These licence surrender rates are used to forecast licence hold-ng by different age groups to the year 2030, and hence the numbersf drivers in different age groups. Fig. 3a and b shows the forecasticence holding to year 2030. For all age groups over forty, licenceolding is based on following cohorts who are already qualified toold licences, For ages up to forty, at least some of the forecast isased on smooth extrapolations of current trends. It does appearhat for teenagers, licence holding will be the same for men andomen.
Fig. 3a shows that by 2030, licence holding by men aged 40 to0 will be steady at around 90% and for men aged 70 and over itill be steady at about 80%. Men aged 20–40 will have lower levels
f licence holding, because of the reduced holding by teenagers.For women aged 30–70, licence holding will have stabilized at
bout 78%. Licence holding by women aged 70 and over will still beising, as will that by women aged 21–29.
. Car driver safety
.1. Trend in car driver casualties
Data on road accident casualties in Britain are published annu-lly by the Department for Transport as Reported Road Casualtiesreat Britain (Department for Transport, annual (b)). This is basedn police reports of road accidents in which someone is injured,
ncluding slight injuries. There is no available record of damage-nly accidents.For most age groups, car driver casualties of all severities (thats, the total of all injuries and deaths) peaked at some point in the
Fig. 3. (a) Forecast holding of car driving licences by men in Britain. (b) Forecastholding of car driving licences by women in Britain.
1990s and are now falling. For drivers aged 60 and over, the fall isvery small, and for those aged 80 and over, the number is roughlysteady (Fig. 4a).
Car drivers aged 30–39 suffered about 200 deaths per yearbetween 1980 and 2004, and teenage drivers about 150 deaths peryear over the same period (Fig. 4b). For drivers aged 40–70, thenumber of deaths have declined fairly steadily since 1980, and forthose aged 70–79, since 1995 (the first year data for this age groupare available). Deaths of drivers aged 80 and over have been steadyat about 60 a year, although now falling. For all drivers aged 70 andover, the number of deaths increased until 1990, remained steadyat about 160 per year until the late 1990s, and has subsequentlyreduced about 40%. Deaths of drivers aged 20–29 have been falling,but for much of the time since 1975 have been over 300 per year.
3.2. Trends in car driver casualty rates per driver licence
These data can be presented as a casualty rate per driving licence(Fig. 5a), which portrays the risk of accident or injury per year. Forall drivers aged 60 and over, the total casualty rate has declinedsteadily since 1980. For younger drivers, the reduction only startedafter year 2000. For most of the period for which data are available,drivers aged 60–69 had the lowest casualty rate per driver licence.
Fatality rates for car drivers have fallen steadily for all driversaged 30 and over (Fig. 5b). For younger drivers, the reduction inthe fatality rates started more recently. The reductions have beengreatest for the oldest drivers. Thus, the reduction for 60- to 69-year olds is steeper than for younger drivers, even greater for 70-
to 79-year old drivers and steepest of all for drivers aged 80 andover. Since 2006, the reduction for teenage drivers has been veryrapid.
736 C.G.B. (Kit) Mitchell / Accident Analysis and Prevention 50 (2013) 732– 741
0
5
10
15
20
25
30
35
40
45a
b
20102005200019951990198519801975
Nu
mb
er
of ca
su
altie
s, th
ou
sa
nd
s
16 to 19
20 to 29
30 to 39
40 to 49
50 to 59
60 to 69
70 to 79
80+
70+
Age group
0
50
100
150
200
250
300
350
400
450
500
20102005200019951990198519801975
De
ath
s
16 to 19
20 to 29
30 to 39
40 to 49
60 to 69
70 to 79
80+
70+
Age group
Fig. 4. (a) Number of car driver casualties (injuries and deaths) for various agegroups, Reported Road Casualties Great Britain (Department for Transport, annual(b)) Britain, Reported Road Casualties Great Britain (Department for Transport,ag
rfboksrit
3
doTN
cc
pFdoTdo
0
200
400
600
800
1000
1200
1400a
b
1975 198 0 198 5 199 0 199 5 200 0 200 5 2010
Ca
su
altie
s p
er
10
0,0
00
lic
en
ce
s
16 to 19
20 to 29
30 to 39
40 to 49
60 to 69
70 to 79
80+
70+
Age group
0
2
4
6
8
10
12
14
16
18
197 5 1980 19 85 1990 1995 2000 2005 201 0
De
ath
s p
er
10
0,0
00
drivin
g lic
en
ce
s
16 to 19
20 to 29
30 to 39
50 to 59
60 to 69
70 to 79
80+
70+
Age group
Fig. 5. (a) Casualty rate for car drivers per 100,000 driver licences. Britain (all injuriesand deaths) (age group 50–59 omitted for clarity). (b) Fatality rate for car driversper 100,000 driver licences – Britain (age group 40–49 omitted for clarity).
0
500
1000
1500
2000
2500a
b
1994 19 96 1998 2000 2002 20 04 20 06 2008 2010All
ca
su
altie
s p
er
10
00
mill
ion
mile
s d
rive
n
16 to 19
20 to 29
30 to 39
40 to 49
50 to 59
60 to 69
70+
Age group
0
5
10
15
20
25
30
35
199 4 1996 19 98 2000 2002 2004 2006 20 08 201 0
De
ath
s p
er
10
00
mill
ion
mile
s d
rive
n
16 to 19
20 to 29
30 to 39
50 to 59
60 to 69
70+
Age group
nnual (b)). (b) Number of car driver deaths for various age groups, Britain (ageroup 50–59 omitted for clarity) for various age groups, Britain.
Thus, over the period since 1996, the oldest drivers have beeneducing their casualty rates fastest. Since 2004, the reduction inatality rate for drivers aged 80 and over, and 70 and over, haseen sufficient to more than balance the increase in the numberf older drivers. The reason for this reduction in fatalities is notnown, though it has been suggested that many drivers now in theireventies have been driving since teenage, and that this early expe-ience may enable them to better compensate for the inevitablencreases in reaction time and loss of visual performance and abilityo multi-task associated with ageing.
.3. Trends in car driver casualty rates per mile driven
Another measure of risk for drivers is the casualty rate per mileriven. Data on the miles driven by car and van drivers by age arenly available for a few years and only since the 1995/1997 Nationalravel Survey. The age groups are the standard ones used by theTS, so that older drivers are in age groups 60–69 and 70 and over.
The data that are available have been used to calculate the totalasualty rate for car drivers (Fig. 6a) and the fatal casualty rate forar drivers (Fig. 6b) per 1000 million miles driven.
Because the miles driven per year are not changing rapidly, theatterns in Fig. 6a and b are fairly similar to those in Fig. 5a and b.ig. 6a shows that since 1996 the car driver casualty rate per mileriven has been reducing steadily for all driver groups aged fifty andver. For younger drivers, reductions have started more recently.
he casualty rate per mile is lowest for drivers aged 60–69. Forrivers aged 70 and over, the casualty rate per mile is between thatf the groups 30–39 and 40–49.Fig. 6. (a) Casualty rate for car drivers per 1000 million miles driven. Britain (allinjuries and deaths). (b) Fatality rate for car drivers per 1000 million miles driven –Britain (age group 40–49 omitted for clarity).
C.G.B. (Kit) Mitchell / Accident Analysis a
10
100
1000
10000
1980 1990 2000 2010 2020 2030
De
ath
s p
er
10
,00
0,0
00
lic
en
ce
s
50-5 9
60-6 9
70-7 9
80+
Age group
Ft(
mp6at
3
aisrd
fci
nittcupphb
ig. 7. Car driver fatality rates per licence for older driver groups with extrapola-ion to 2030 for various age groups, Britain, Reported Road Casualties Great BritainDepartment for Transport, annual (b)).
Fig. 6b shows the remarkable reduction in the fatality rate perile for car drivers aged 70 and over. In addition, the fatality rates
er mile have been falling since 1996 for drivers aged 50–59 and0–69, and more recently for younger drivers. The rate for driversged 70 and over is high compared to other age groups, excepteenagers, because of the effects of fragility.
.4. Forecasting car driver casualties
When casualty rates are plotted on log-linear paper, trends oftenppear as straight lines. Fig. 7 shows a plot of this kind for fatal-ty rates for car drivers of various ages over 50 years old. Thishows how the best fit lines can be extrapolated, to estimate futureates, on the basis that current trends continue. Again, the steepestecline is for the age group 80 and over.
These extrapolated fatality rates can be applied to the forecastuture population of licence holders, to determine the number ofasualties in the future. When this is done, the results are as shownn Fig. 8.
The forecast shows that, except for the 16–19 age group,umbers of fatalities are likely to continue to fall. However,
mprovements in road safety since 2005 have been so great thathe current numbers of casualties are well below the long-termrend lines. This makes it very difficult to extrapolate, because aontinuation of the trend lines require casualty numbers to jumpp in ways that they are certainly not expected to do. What could
ossibly happen, though it is hoped not, is that fatality rates couldlateau for the next ten years or so, until the long term trend linesave come down to the current values. This happened in Britainetween 1993 and 2003.0
50
100
150
200
250
300
350
400
450
203020202010200019901980
Num
ber
of car
driver
fata
litie
s
16-19
20-29
30-39
40-49
70-79
80+
70+
Age group
Fig. 8. Forecast car driver fatalities to year 2030.
nd Prevention 50 (2013) 732– 741 737
In particular, the number of older driver fatalities will continueto fall. This is a result of the fatality rate per driver for older driversreducing faster than their number is increasing.
4. Older driver safety in detail
Much research has examined the safety of older drivers, aspectsof driver behaviour that change as people age, how drivers regulatetheir driving and try to compensate for problems that develop asthey age, and the effectiveness of programmes to improve the safetyof older drivers.
In the 1990s this research initially focused on the car driverfatality rate per mile driven, which increases sharply with age aftermiddle age (TRB, 1988). Further research has shown that much ofthe increase in driver fatalities in older age is a result of the way thatpeople become physically more fragile as they age (see Section 5).This means that for older road users, a given accident is more likelyto cause injury, and a given injury is more likely to cause death.
There is no doubt that competences such as vision, reaction timeand the ability to multi-task deteriorate with increasing age, andthat some older drivers have problems with driving that mean thatthey can no longer continue to do so safely. Older drivers drivedifferently to younger drivers, and regulate their driving to avoidconditions that they find difficult or stressful. But much analysishas also shown that on average the accident involvement rate perdriver does not increase until ages over seventy-five or eighty. Also,much of any increase in accident involvement is a consequence ofthe reduced mileage driven by older drivers – the ‘mileage bias’. Arecent study shows that only those who drive less than 3000 km(1860 miles) per year show an increase in accidents per distancedriven after age 75 (Langford et al., 2006).
4.1. Accident involvement
The Department for Transport publication ‘Reported Road Casu-alties Great Britain’ (Department for Transport, annual (a)), whichuses data from police records of injury accidents to tabulate casu-alties by injury level (fatal, serious and slight) and by many factorssuch as type of road user, casualty age and class of road, makespossible analyses of car driver casualties of various severities bydriver age. The National Travel Survey (Department for Transport,annual (b)), an annual household travel survey of some 20,000 per-sons, provides data on driver licensing and on travel as a car driver.As with any travel survey, there are questions over possible under-reporting of travel in a seven-day travel diary, but great care is takenin NTS to obtain as complete as possible record of travel under-taken. Also, there is no reason to believe that any under-reportingis greater for some age groups than others. The survey makes it pos-sible to calculate the casualty rate per driver licence and per miledriven. Fig. 9a shows the casualty rates per driver licence for fatal,killed and seriously injured casualties (KSI), slight injuries and allcasualties (that is, all deaths and injuries) in Britain in 2010.
The casualty rate for slight injuries is probably the best proxymeasure of accident involvement as it is the most free from biascaused by the increase in fragility with increasing age. There aremany more damage-only accidents than slight injury accidents,but data on damage-only accidents are not available. It is likelythat variation with age in the rate of involvement in all accidents(including damage only accidents), is closer to that for slight injuryaccidents than for accidents that cause more serious injuries.
Fig. 9a shows how much of the increase in fatal and serious casu-
alties is a result of fragility. The increase in fragility with age causesthe curve for fatalities, and to a lesser extent for KSI casualties, tobend up for older ages, even though the curve for slight injuriesis almost flat. The rate for slight injuries is lowest for drivers aged
738 C.G.B. (Kit) Mitchell / Accident Analysis and Prevention 50 (2013) 732– 741
0
100
200
300
400
500
600
700
800
900a
b
908070605040302010
Driver age
Ca
su
alty r
ate
s p
er
drive
r
Deaths per 10 milli on driv ers
KSI casual ties per mi lli on drive rs
All cas ualties pe r 100,0 00 driv ers
Sli ght in juries per 100,000 drive rs
0
500
1000
1500
2000
2500
3000
3500
908070605040302010
Driver age
Ca
su
alty r
ate
s p
er
dis
tan
ce
drive
n
Deaths per 100 bil lio n mi les
KSI cas ual ties per 10 billion mile s
All casual ties per bi lli on mile s
Slight injuries per billi on mil es
Fig. 9. (a) Car driver casualty rates per driving licence–Britain, 2010. Reported RoadCasualties Great Britain (Department for Transport, annual (b)) and National TravelSurvey (Department for Transport, annual (a)). (b) Car driver casualty rates per dis-tT(
6ffiu
Bi6riad
i(ihwif
4
aop
0
500
1000
1500
2000
2500
201020082006200420022000199819961994Slig
ht ca
su
altie
s p
er
10
00
mill
ion
mile
s d
rive
n
16 to 19
20 to 29
30 to 39
40 to 49
50 to 59
60 to 69
70+
Age group
a similar result.When the British data are converted to show the number of
pedestrians killed per driver in each age group (Fig. 11b), the risk
0
10
20
30
40
50
60a
b
10 20 30 40 50 60 70 80 90
Age of car driver
Pe
de
str
ian
s k
ille
d
Tot al
Male driver s
Female driv ers
0
10
20
30
40
50
60
70
80
90
100
10 20 30 40 50 60 70 80 90
Age of car driver
Pe
de
str
ian
s k
ille
d p
er
mill
ion
drive
rs
Male driver s
Female drivers
Total
ance driven–Britain, 2010. Reported Road Casualties Great Britain (Department forransport, annual (b)) and National Travel Survey (Department for Transport, annuala)).
0–79, and increases by 20% for drivers aged 80 and over. The curveor fatalities is lowest for age 50–59, and increases by a factor of sixor drivers aged 80 and over. This increase is larger than normal;n previous years since 2005, the increase in the fatality rate hassually been by a factor of about four.
Fig. 9b shows the casualty rates per distance driven, also forritain in 2010. As expected, there are larger increases in rates with
ncreasing age. The age for the minimum rate for slight injuries is0–69, and the increase in the slight injury rate for ages 70–79elative to that minimum is 27%. The rate for ages 80 and overs significantly higher, about two and a half times the value forges 60–69. But because this age group average much less distanceriven, their casualty rate per year (Fig. 9a) only increases 20%.
The trend in accident involvement per mile driven can bendicated by plotting the rate of slight injury casualties per mileFig. 10). This shows the steady reduction in the rate of slightnjuries per mile for car drivers aged over 60. The age group 60–69as improved faster than that aged 50–59, to become the groupith the lowest casualty rate. The group aged 70 and over has
mproved even more rapidly, to become comparable to the rateor the group aged 40–49.
.2. Danger to other road users
Data on the characteristics of drivers at fault in accidents are notvailable in Britain on a national basis. A good alternative measuref the danger that drivers pose to other road users is the number ofedestrians killed by drivers of various ages. Even if the pedestrian
Fig. 10. The trend in car driver slight injury casualties per 1000 million miles driven.Reported Road Casualties Great Britain (Department for Transport, annual (b)) andNational Travel Survey (Department for Transport, annual (a)).
is blameworthy in some accidents in which a pedestrian is killed,the number of pedestrians killed is a measure of the extent to whicha group of drivers tends, on average, to drive defensively, preparedto react to unexpected incidents.
Data for Britain in 2005 show that the majority of pedestrianswho are killed by cars are killed by young or middle aged maledrivers (Fig. 11a) (special tabulation by Department for Transport).Relatively few pedestrians are killed by drivers aged 60 and over. Asimilar analysis by Hakamies-Blomqvist (2004) for the USA shows
Fig. 11. (a) Number of pedestrians killed by cars with drivers of different ages.Britain 2005 Special tabulation by the Department for Transport from police reportsof accidents. (b) Risk of a car driver killing a pedestrians by driver age and genderBritain 2005.
C.G.B. (Kit) Mitchell / Accident Analysis a
0
1
2
3
4
5
6
7
8
9080706050403020100
Age
Pe
rce
nt o
f ca
su
altie
s w
ho
die
Car drivers
Car passengers
Pedestrians
FB
oddad
5
oii2
apdff
faafispcp
F2
ig. 12. Fragility – the percentage of people injured in traffic accidents who dieritain 2006–2009.
f a pedestrian being killed per million drivers shows that youngrivers have a much higher risk of killing a pedestrian than dorivers aged over thirty. Drivers aged 60–75 pose the lowest risk ofll drivers, but over age 75 there is an increase in risk for the oldestrivers.
. Fragility
As has been mentioned earlier in this paper, as people growlder, they become more physically fragile. That is, if they arenvolved in a given accident they are more likely to be injured, andf injured, they are more likely to die from those injuries (Evans,000).
Because of the increase in fragility with age, older road usersre more likely to die if they are involved in a traffic accident. Foredestrians aged 30–50, about 2% of casualties of traffic accidentsie from their injuries. For those aged 60–70 the percentage is 3.6%;or 70–80, 5.8% and for 80 and over, 8.6% (Fig. 12). The percentagesor car occupants are lower, but the pattern is similar.
The physical fragility of the person injured can be separatedrom the degree of protection provided by a vehicle by deriving
‘Fragility Index’. This is the percentage of injuries that are fatal for particular road user divided by the percentage of injuries that areatal for that type of road user at a particular age, often 30–39. Thiss shown for pedestrians and car occupants in Fig. 13, and it will be
een that the variation in Fragility Index with age is very similar foredestrians, car drivers and car passengers. The index is higher forar occupants aged 17–29 than for middle age occupants, and thisrobably reflects the greater severity of the accidents that young0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
9080706050403020100
Age
Fra
gili
ty r
ela
tive
to
ag
e 3
0 -
39
Car drivers
Car passengers
Pedestrians
ig. 13. Fragility Index – fragility relative to that for people aged 30–39 Britain006–2009.
nd Prevention 50 (2013) 732– 741 739
adult car occupants experience, which means that their chance ofsurvival is lower.
6. Discussion
6.1. Trends in driving licence holding
The trend in licence holding shown in Section 2 is typical oftrends in other countries for which data are available. It is the resultof people, once they have obtained a driver licence, retaining it intoold age.
Until about 1990, as the country became more prosperous, moreteenagers obtained car driving licences and then carried themthrough middle age into older age. After 1992 the percentage ofteenagers obtaining car driving licences started to drop, falling toa low in 2004, after which it has started to recover. The reason forthis drop has not been studied in detail, but a factor must be theincreasing costs of learning to drive and of insuring and operatinga car.
The estimates of the percentage of licence holders who ceaseto hold a licence in older age is believed to be new, made possibleby the growing period of time series data from the National TravelSurvey. The estimate makes it possible to predict future levels oflicence holding by older people, as described in Section 2.4. It isinteresting that even by the year 2030, licence holding by womenaged 70 and over will still be rising, though for men of this agelicence holding will have stabilised.
Data on the number of older drivers whose licences are with-drawn are not published, but anecdotal reports suggest that veryfew are withdrawn by the licensing authority as opposed to beingsurrendered voluntarily or allowed to lapse when they come up forrenewal.
6.2. Trends in casualties
The reduction in casualty and fatality numbers and rates forolder drivers since the mid-1990s in Britain is similar the trendin USA and a number of other countries. In Sweden the number ofdeaths of car drivers aged 75 and over averaged 38 for 1999–2002and had fallen to 11 in 2010 (Trafikanalys, annual). In USA the num-ber of deaths of car drivers aged 75 was 2030 for 1995 and 1996and had fallen to 1733 in 2009 (FARS, annual). In Britain the num-ber of deaths of car drivers aged 70–79 was 87 averaged over 1995and 1996 and had fallen to 39 in 2010. The corresponding figuresfor drivers aged 80 and over were 60 for 1995 and 1996, falling to43 in 2009 but increasing again to 58 in 2010. The correspondingfatality rates per million licences for 2009 are 126 in USA, 34 inSweden and in Britain, 39 for drivers aged 80 and over and 20 forthose aged 70–79.
Not only have the absolute number of deaths of older driversfallen, but the percentage of all driver deaths represented by theoldest driver age group has generally been falling. In Swedendrivers aged 75 and over averaged 14.3% of all car driver deathsfor 1999–2002, which fell to 10.1% of all deaths in 2010. In USA thesame age group averaged 8.5% of all driver deaths in 1995 and 1996,which fell to 7.5% for 2006–2008 but increased to 8.0% in 2009. InBritain car drivers aged 70–79 averaged 7.8% of all car driver deathsin 1995 and 1996, which fell to 6.8% in 2010. However, drivers aged80 and over averaged 5.4% of all car driver deaths in 1995 and 1996,which had become 6.1% for the period 2006–2009 and increasedfurther to 10.1% in 2010. For this age group the absolute number of
deaths per year is only 50–60 a year, so statistical variability is tobe expected.This reduction in the absolute number of deaths of older cardrivers means that the risk of death per driver is reducing faster
7 lysis a
taitowiaawidaid
aadiIahthpohsOwr
iit2tuwjIi
7
s
•
•
•
•
•
40 C.G.B. (Kit) Mitchell / Accident Ana
han the number of drivers is increasing. It would be good to beble to explain why this is happening, but there have been no stud-es that provide convincing explanations. In Britain and Swedenhere have been no changes to the process of licence renewal forlder drivers, so there is no reason to suppose that older driversith problems are stopping driving at an increased rate. Indeed,
n Sweden the percentage of people aged 75 and over who held car driving licence increased from 72.0 to 76.3% between 2008nd 2010. In Britain the percentage of people aged 80 and overho held a car driving licence increased from 29% in 2002 to 38.3%
n 2008 and 41.7% in 2010 The reduction in the percentage ofriver deaths represented by the oldest drivers in Sweden and USA,nd drivers aged 70–79 in Britain, means that the risk of fatal-ty for these age groups is reducing faster than the average for allrivers.
The reduction in fatalities for older car drivers is happening at time when fatalities are reducing rapidly for all road users of allges in a number of countries. These reductions must be in partue to packages of road safety measures, but in many countries it
s difficult to identify which measures have led to improvements.n some, such as Sweden, France and Spain, the successful pack-ges can be identified. Sweden, under the heading ‘Vision zero’,as improved infrastructure and reduced speed limits to makehe road system more forgiving of driver error. France and Spainave each introduced packages consisting of extensive overt trafficolicing, enforcement of drink-drive legislation and enforcementf speed limits, usually by means of cameras. In addition, Spainas built about 11,000 km of high quality dual carriageway roadsince 1980, to take traffic off dangerous single carriageway roads.ver the same period France has built about 7000 km of motor-ays and improved large numbers of junctions on all classes of
oads.Britain has made extensive use of cameras to enforce speed lim-
ts. As a result, the percentage of cars exceeding the 30 mph limitn urban areas has reduced from 69% in 1998 to 46% in 2010, andhose exceeding 35mph in 30mph areas from 32% in 2001 to 16% in010 (Department for Transport, 2011; Mitchell, 2012). It is likelyhat reducing the number of vehicles travelling excessively fast inrban areas provides a disproportionate benefit for older drivers,ho have a higher than average percentage of their accidents at
unctions on urban roads. However, this is unproven speculation.t is necessary to recognise that the causes of recent improvementsn road safety have not yet been explained in detail.
. Conclusions
Data for Britain from police reports of all accidents in whichomeone is killed or injured show that:
Car drivers up to age 75 or 80 are, on average, no more likely to beinvolved in a slight injury accident than younger drivers on thebasis of casualties per driving licence per year.The rate of car driver casualties per mile driven does increase forages over 70, and increases by almost a factor of three for driversaged 80 and over.Corresponding fatality rates rise much more rapidly with increas-ing age, because of the physical fragility of older people whichincreases with age.Car drivers aged sixty and over kill many fewer pedestrians thando younger drivers, and this is an indication that older drivers do
not pose an excessive risk to other road users. Data from FARSshow a similar result for USA.Older drivers are at greater risk of injury or death themselves,because of their greater physical fragility that results from ageing.nd Prevention 50 (2013) 732– 741
• The number of deaths of older car drivers has been reducingsince 2004, despite the increase in number of older drivers.This applies both to drivers aged 80 and over, and thoseaged 70 to 79. Similar trends are occurring in USA and inSweden.
• The reason for the reduction in older driver fatalities and the fatal-ity rate per driver licence in Britain is not known. There is noevidence that screening to identify drivers who should no longerbe driving has improved.
Data from the British National Travel Survey show that:
• Driving licence holding is continuing to increase for older drivers,and particularly for older women drivers.
• Licence holding by men aged 40–70 and by women aged 30–60has stabilized.
• Licence surrender begins to occur after age seventy. By age 80,10% of women who held a licence at age 70 have surrendered it,as have 5% of men. By age 90, 38% of women and 25% of men whohad held a licence at 70 have surrendered their licences.
• A forecast of older driver casualties shows that the number ofolder driver fatalities is likely to continue to fall. In absoluteterms, the number of older driver fatalities will be relativelysmall.
• While there are undoubtedly some older drivers who shouldstop diving for safety reasons, these results suggest that thesafety of older drivers in Britain is currently being managed ina way that should be sustainable for at least the next twentyyears.
• The licence renewal process outlined in Section 2.2 appears toprovide a satisfactory balance between mobility and safety, eventhough it relies on the driver to self-report medical problems tothe licensing authority.
References
Box, G., Mitchell, 2010. Maintaining Safe Mobility for the Ageing Population. RoyalAutomobile Club Foundation for Motoring, London.
Burkhardt, J.E., McGavock, A.T., 1999. Tomorrows older drivers: Who? How many?What impacts? Paper 99-0934. In: Transportation Research Board 78th AnnualMeeting, 10–14 January 1999, Washington, DC.
Department for Transport, annual (a). National Travel Survey Department for Trans-port, London.
Department for Transport, annual (b). Reported Road Casualties Great BritainDepartment for Transport, London (Previously Road Casualties Great Britain andRoad Accidents Great Britain).
Department for Transport, 2011. Free Flow Vehicle Speeds in Great Britain 2010Statistical Release. Department for Transport, London.
Eby, D.W., Molnar, L.J., St. Louis, R.M. M-CASTL 2008 Synthesis Report: Vol. 1.Older Adult Mobility Report No. M-CASTL-2008-01, University of Michigan.Transportation Research Institute & Michigan Center for Advancing Safe Trans-portation throughout the Lifespan, Ann Arbor, Michigan.
Eustace, D., Wei, H., 2010. The role of driver age and gender in motor vehicle fatalcrashes. Journal of Transportation Safety & Security 2 (March (1)), 28–44.
Evans, L., 2000. Risks older drivers face themselves and threats they pose to otherroad users. International Journal of Epidemiology 29, 315–322.
FARS, annual. Traffic Safety Facts A Compilation of Motor Vehicle Crash Datafrom the Fatality Analysis Reporting System, National Highway TrafficSafety Administration, US Department of Transportation, Washington DC20590.
Hakamies-Blomqvist, L., Johansson, K., Lundberg, C., 1996. Medical screeningof older drivers as a traffic safety measure—a comparative Finnish-Swedishevaluation study. Journal of the American Geriatrics Society 446 (June),650–653.
Hakamies-Blomqvist, L., 2004. Safety of Older Persons in Traffic. In: ‘Transportationin an Aging Society—A Decade of Experience’, Transportation Research BoardConference Proceeding, No. 27, Washington, DC, pp. 22–35.
Langford, J., Fitzharris, M., Koppell, S., Newstead, S., 2004. Effectiveness of manda-
tory license testing for older drivers in reducing crash risk among urban olderAustralian drivers. Traffic Injury Prevention 5 (4), 326–335.Langford, J., Methorst, R., Hakamies-Blomqvist, L., 2006. Older drivers do not have ahigh crash risk—a replication of low mileage bias. Accident Analysis & Prevention38 (May (3)), 574–578.
lysis a
M
M
N
C.G.B. (Kit) Mitchell / Accident Ana
itchell, C.G.B., 2008. The licensing of older drivers in Europe—a case study. TrafficInjury Prevention 9 (4), 360–366.
itchell, C.G.B., 2012. Speed and Safety: Evidence from Published Data. Royal Auto-mobile Club Foundation for Motoring, London.
HTSA, annual. Traffic Safety Facts—A Compilation of Motor Vehicle Crash Datafrom the Fatality Analysis Reporting System and the General Estimates Sys-tem National Highway Traffic Safety Administration, National Center for
nd Prevention 50 (2013) 732– 741 741
Statistics and Analysis, U.S. Department of Transportation, Washington, DC20590.
Trafikanalys, annual. Vägtrafikskador (Road traffic injuries) Statistik 2011:15, Trafik-analys, Stockholm.
TRB (1988) Transportation in an Aging Society—Improving Mobility and Safety forOlder Persons Special Report 218, Transportation Research Board, Washington,DC.