Written evidence from Professor Richard
Allsop (DDD 09)
I am submitting this evidence as an individual
expert, not on behalf of UCL.
I provided in evidence to Sir Peter North
the more cautious of the two estimates he quotes of numbers of
deaths and serious injuries that would be prevented by reducing
the permitted blood alcohol concentration (BAC) as he proposes.
I have followed developments in drink driving law in Britain since
the mid-1960s, when I provided as a Scientific Officer of the
then Road Research Laboratory one of the main statistical analyses
on which the introduction of the BAC limit in 1967 was based.
I am a director of PACTS and of the European Transport Safety
Council.
SUMMARY
Except for concluding remarks which relate also
to drug driving, my evidence is confined to drink driving law.
1. The permitted limit on blood alcohol
concentration (BAC) should definitely be reduced as proposed,
primarily because the substantial numbers of deaths and serous
injuries that this will prevent would clearly outweigh any foreseeable
adverse effects of the change, and it is premature to consider
a greater reduction than is proposed. The range of estimates of
these numbers mentioned in the Review Report is discussed in some
detail in this evidence.
2. The mandatory one year driving ban is
appropriate for offenders at the new lower level, but if its severity
were the principal obstacle to public and parliamentary acceptance
of the reduced limit, a lower penalty based on "two strikes
and you're out" would be an acceptable alternative for offences
at BACs between 50 and 80mg/100ml only.
3. Reducing the permitted BAC limit would
have secondary benefits in terms of consistency with limits in
other European countries and with the message "Don't drink
and drive", and might make a small contribution to wider
efforts to moderate the use of alcohol. It would challenge establishments
serving alcoholic drinks and to which customers travel by car
to offer attractive alternative drinks to the drivers among their
customers.
4. Policy and practice in enforcement of
the BAC limit would need to make it clear to drivers that those
exceeding the reduced limit were at real risk of detection and
conviction and that those with BACs over 80mg/100ml were at least
as likely to be detected as before. The police will be greatly
helped in this by urgent completion of the already long overdue
type approval of evidential roadside breathtesting equipment,
removal of the "statutory option", inclusion of drink
driving among the "Offences Brought to Justice" on which
police forces are required to report, and appropriately widened
powers to breathtest drivers.
5. Drink and drug driving law should be
focussed on road safety rather than on wider policy on alcohol
and other drugs, and should reflect the degree of risk caused
by impairment. For drink, the risk is closely related to the level
of alcohol in the blood, but this relationship is less clear for
other drugs, for which impairment, not just presence, needs to
be established.
Should the Permitted Blood Alcohol Limit be reduced
as proposed?
1. The limit should definitely be reduced
as proposed, primarily because this can be expected to prevent
substantial numbers of deaths and serous injuries in road traffic
accidents by complementing existing and further efforts to deter
driving at very high BACs with a parallel effort to deter driving
at lower BACs that are nevertheless associated with substantially
increased risk. It is premature to consider a greater reduction
in the limit than is proposed.
2. Sir Peter's report summarises in paragraphs
3.40-3.44 a range of estimates of numbers of deaths and serious
injuries that could be prevented—estimates made by researchers
at the University of Sheffield and in my own written evidence
to the Review. These estimates are discussed in paragraphs 4.15-4.18
in the course of the argument leading to Recommendation (3) to
reduce the BAC limit from 80 to 50 mg/100ml. Although the estimates
differ widely, the Review concludes that they all indicate that
substantial numbers of deaths and serious injuries would be prevented
each year.
3. The report provides a reference to a
report from the Sheffield researchers giving full details of the
basis for their estimates, but for my own estimates there is no
such reference. In case it helps the Committee, I therefore attach
for reference a note prepared in 2008 for the Department for Transport
(DfT) giving similar details for my own estimates. In view of
the considerable difference between our two sets of estimates,
especially in numbers of deaths prevented, I have, since seeing
the Sheffield estimates for the first time in Sir Peter's report,
investigated just why and how the estimates differ so widely.
What the two methods have in common and where they differ are
summarised in the Appendix to this evidence, and the differences
in method make it perfectly understandable that the resulting
estimates themselves differ.
4. Just how they differ can be illustrated
by applying both methods to deaths in 2007—the latest year
for which all the required data were available when evidence was
submitted to the North Review. This leads to the following distributions
among the BAC ranges 0-20, 20-50, 50-80, 80-110 and over 110mg/100ml
of the deaths that it is estimated would have been prevented in
Great Britain in 2007 according to the two methods, and for the
Sheffield method according to how long before 2007 the reduced
limit had been in force.
BAC of drinking driver* (mg/100ml)
| Deaths prevented in 2007 as estimated by Sheffield approach with limit reduced on
| | | |
|
1 January 2007
|
1 July 2005 |
1 July 2001
| Allsop
approach. |
| (A) | (B)
| (C) | (D) |
110 and over | 102
| 247 | 350 | 0
|
| 80-110 | 12 | 29
| 41 | 17 |
| 50-80 | 3 | 13
| 29 | 13 |
| 20-50 | 11 | 5
| 40 | 17 |
| 0-20 | 0 | -4
| -1 | 0 |
| All BAC levels | 128 | 290
| 459 | 47 |
* under the existing 80mg/100ml limit
| | | |
|
| |
| | |
Comparing Columns A, B and D shows that the two approaches
are found to lead to not dissimilar estimates of the numbers of
deaths prevented in the early years of the reduced limit in accidents
involving drivers with BACs below 110mg/100ml. The bulk of the
difference in the estimates given in these three columns by the
two approaches lies in the estimated prevention of deaths in accidents
where a driver has a BAC of over 110mg/100ml.
5. It is clear how this comes about, because my own estimates
exclude the possibility of preventing deaths in accidents where
a driver has a BAC of over 110mg/100ml by reducing the BAC limit.
I did this on the grounds that so much has already been done since
1967 to reduce such deaths in Britain, where only an estimated
0.3% of all driving now takes place at BACs over 80mg/100ml. The
Sheffield researchers, on the other hand, base their estimate
on the effect of reducing the limit on drivers at these BACs being
similar to that achieved in Australia, where 2.5% of driving was
previously taking place at BACs over 80mg/100ml.
6. While it is not unreasonable in the light of the Australian
and other overseas experience to hope for some prevention of deaths
at high BACs as a result of increased public information and enforcement
associated with a reduction in the limit, I think we in Britain
should be cautious in our expectations in this respect because
of what has already been achieved here in respect of deaths at
high BACs.
7. The Sheffield researchers estimate further, again
on the basis of the Australian experience, that the numbers of
deaths prevented would increase in subsequent years, as illustrated
in Column C of the above table, which estimates what the effect
would have been in 2007 if the limit had been reduced six years
previously. My own estimates do not consider this possibility,
but Column C would imply that seven out of eight of the deaths
in 2007 in accidents involving drivers with BACs greater than
20mg/100ml would have been prevented. This would be wonderful,
but seems a good deal too much to expect. Another possibility
considered by the Sheffield researchers is, however, that the
effect of reducing the limit here might be half what it was in
Australia, and it is worth noting that halving the numbers in
Column C of the table brings those for BACs below 110mg/100ml
roughly into line with those in Column D.
8. Although this discussion is necessarily rather detailed,
I hope the Committee may find it useful in helping interpret the
range of estimates presented in evidence to the North Review.
Is the mandatory One-year Driving Ban appropriate for less
severe offenders under the lower BAC limit of 50mg/100ml?
9. In the argument leading to Recommendation (15) for
such a ban, Sir Peter reports receiving very persuasive evidence
in favour of this level of penalty, and on that basis I am also
convinced that it is appropriate.
10. If, however, as implementation of the Review's recommendations
is considered, the severity of this penalty were to become the
principal obstacle to public and parliamentary acceptance of the
reduced limit, an acceptable alternative could be a basic penalty
for driving with a BAC between 50 and 80mg/100ml of six penalty
points remaining on the licence for 10 years together with up
to three months imprisonment and a fine of up to £2,500.
This would mean "two strikes within 10 years and you're out".
This would be on the understanding that the penalties for exceeding
80mg/100ml and for failing to provide a specimen should remain
as they are now.
What wider costs and benefits are likely to result from changes
to drink driving law?
11. As well as preventing deaths and serious injuries,
lowering the BAC limit will also bring the advantages of:
— reducing drivers' margin for discretion in relation
to the advice not to drink alcohol at all before driving,
— better preparing the millions of British drivers who
drive in mainland Europe to comply with the law there,
— removing any temptation to drivers coming to Britain
from mainland Europe to drink more before driving while they are
here because our limit is known to be so high, and
— removing the demoralising effect on those working
at all levels on this and other aspects of road casualty reduction,
and the damaging effect on the public perception of Government
commitment to casualty reduction, of knowing that an important
means of reducing death and injury, widely approved by the public,
continues to be neglected by Government.
It may also make a small contribution to wider efforts to
moderate the use of alcohol, but this should not be seen as an
objective in itself of a change in drink driving law.
12. It is reasonable to recognise that reducing the BAC
limit might well lead to some reduction in driving to pubs, clubs
and restaurants by patrons who are not prepared to adapt their
lifestyles to the extent of one driver per vehicle restraining
themselves to the level of consumption of alcohol needed to be
below the reduced limit for the return journey. To that extent
the relevant trade may well be affected, and the businesses concerned
will be challenged by this as they are by other ups and downs
of trading conditions. The fact that they have faced this challenge
for the sake of reducing the numbers killed or seriously injured
on the roads could be kept in mind when other aspects of public
policy that may affect them are being considered, such as the
scope for helping the village pub to accommodate a community shop
and post office.
What would be the Implications of changes in Drink Driving
Law for Enforcement?
13. Policy and practice in enforcement of the reduced
BAC limit would need to make it clear to drivers that those exceeding
the reduced limit were at real risk of detection and conviction
and that those with BACs over 80mg/100ml were at least as likely
to be detected as before.
14. The police can be greatly helped in this in four
main ways:
— by urgent completion of the already long overdue type
approval of evidential roadside breathtesting equipment according
to Recommendation (27);
— by removing the statutory option for a blood or urine
test where a recorded breath alcohol concentration is within a
certain margin above the threshold for prosecution, according
to Recommendation (11), accompanied by implementing Recommendation
(13) and, if the BAC limit is reduced, Recommendations (12) and
(14) (Recommendations (12)-(14) together provide an elegant resolution
of important issues of detail concerning enforcement thresholds
and margins for error in measurement of alcohol concentrations.);
— by inclusion of drink driving involving mandatory
disqualification among the "Offences Brought to Justice"
on which police forces are required to report (Recommendation
(25); and
— by giving them a power to require anyone who is actually
driving to cooperate with a preliminary breath test, which would
allow the police to carry out intelligence-led targeted enforcement
in the course of which any passing driver would be equally likely
to be tested.
15. Enhanced enforcement might well lead to a short-term
increase in workload for the courts and in disruption of offenders'
lives as a result of disqualification, but an emphasis on prevention
through deterrence rather than punishment through detection can
be hoped to result in such impacts being only short-term.
THE PURPOSE
OF DRINK
AND DRUG
DRIVING LAW
16. In order to gain and retain public acceptance of
and respect for drink and drug driving law, it is important that
such law should be seen and understood to be focussed on road
safety, and not be mistaken for an indirect approach to wider
policy and objectives, however desirable in themselves, concerning
alcohol and other drugs. It is also important that offences and
penalties under drink and drug driving law should reflect the
degree of risk caused by impairment. For drink, the risk is closely
related to the level of alcohol in the blood at the time, which
simplifies the definition, detection of offences, and prosecution
of offenders. But this relationship is less clear for other drugs,
which necessitates definition and detection of offences being
based on evidence of drug-related impairment at the time, and
not simply on presence of the drug in the body at the time.
REFERENCESDepartment for Transport
(annually) Reported Road Casualties Great Britain. London:
TSO
Maycock G (1997) Drinking and driving in Great Britain—a
review TRL Report 232 Crowthorne: Transport Research Laboratory
August 2010
APPENDIX
COMMON GROUND AND DIFFERENCES BETWEEN THE SHEFFIELD RESEARCHERS'
METHOD AND ALLSOP'S METHOD OF ESTIMATING NUMBERS OF DEATHS AND
SERIOUS INJURIES PREVENTED
A1. The two sets of estimates have three important starting
points in common:
— Both assume that the amounts by which a driver's risk
of involvement in a fatal accident or in an injury accident increase
as functions of the driver's BAC are as estimated by the two exponential
functions fitted to data for Great Britain by Maycock (1997 Appendix
A).
— Both assume that the distribution of deaths in road
accidents with respect to the BACs of involved drivers is estimated
by the distribution with respect to BAC of killed drivers. This
in turn is estimated each year for Great Britain by the Department
for Transport (DfT) and published in summary form one year in
arrears in Reported Road Casualties Great Britain (RRCGB)
(DfT annually)—previously Road Casualties Great Britain—and
the Sheffield researchers obtained this distribution in detail
for recent years from the DfT. When I made my estimates in August
2008 the distribution seemed to have remained stable for many
years, so I used a quite detailed summary of the same distribution
for 1990-94 from Maycock's (1997) Table 10.
— Both estimate the likely effect of reducing the BAC
limit upon the numbers killed or injured by calculating how these
numbers could be expected to change in response to certain changes
in the distribution of BACs of drivers.
In the absence of any previous experience of changing the
BAC limit in Britain, there is no direct evidence as to how the
distribution of BACs of drivers here might change. The two approaches
differ mainly in their ways of addressing this lack of direct
evidence.
A2. The Sheffield researchers use the Maycock risk function
for fatal accidents in conjunction with the distribution of driver
deaths with respect to BAC to estimate the distribution of driving
with respect to driver BAC. Surveys of this distribution were
made in Adelaide, South Australia just before, and then six months,
two years and six years after, the BAC limit there was reduced
from 80mg/100ml to 50mg/100ml in 1991. The Sheffield researchers
use results of these surveys as the basis for assumptions about
how the BAC distribution of driving in Britain might respond to
the same change in BAC limit here. These assumptions, together
with the pre-existing distribution of driving and the risk functions
lead directly to estimates of reductions in numbers killed and
injured over the first six years after reducing the limit. The
process of estimation is refined by using the estimated BAC distributions
of driving in each of the years 1998-2007 to estimate how this
distribution might itself be evolving over subsequent years in
advance of any reduction in the BAC limit.
A3. My own approach is more rudimentary. It makes what
are intended to be plausible and conservative assumptions about
how those currently driving with BACs in the ranges 20-50, 50-80
and 80-110mg/100ml might respond to the reduction in the BAC limit,
and makes the further conservative assumption that those already
choosing to drive with BACs of over 110mg/100ml would be unaffected
by a change in the limit as such—while recognising that
they might be affected by associated public information and increases
in enforcement activity. These assumptions, together with the
pre-existing BAC distribution of killed drivers and the risk functions,
lead directly to estimates of reductions in numbers killed and
injured in a year. My approach makes no attempt to allow for evolution
of the BAC distribution of driving in advance of any reduction
in the limit or to estimate how the reductions in numbers killed
or injured might change over the years after reducing the limit.
ATTACHMENT
NOTE PROVIDED TO THE DEPARTMENT FOR TRANSPORT THROUGH
TRL LTD BY RICHARD ALLSOP FROM UCL CENTRE FOR TRANSPORT STUDIES
IN AUGUST 2008
NOTE: This note is placed in the public domain with the agreement
of the Department for Transport. In the last paragraph, provisional
estimates from Road Casualties Great Britain 2007 of numbers of
drink-drive casualties in 2007 have been replaced by final estimates
from Reported Road Casualties Great Britain 2008
ESTIMATED CASUALTY REDUCTION IMPACTS OF REDUCING THE DRINK-DRIVE
LIMIT TO BAC 50MG/100ML OR 20MG/100ML
NOTE: All references to alcohol levels in this note are to
blood alcohol concentration (BAC) in units of mg/100ml without
mention of the units.
APPROACH TO
ESTIMATION OF
CASUALTY REDUCTIONS
The approach adopted here in order to estimate casualty reduction
impacts of reducing the drink-drive limit uses information about
numbers killed and injured in road traffic accidents, the distribution
of BACs of drivers killed in road traffic accidents and of drivers
involved in road traffic accidents that result in personal injury,
estimated relationships between accident risk and driver's BAC,
and assumptions about the effect of changes in the drink-drive
limit upon the numbers of accident-involved drivers with various
levels of BAC.
Because of the many relevant differences between countries,
only information about drivers and accidents in Great Britain
and relationships derived from such information are used. The
assumptions about possible effects of changes in the legal BAC
limit upon numbers of accident-involved drivers are developed
by the author from a more limited set of similar assumptions made
by him in an earlier study which has been in the public domain
since 2005.
SOURCES OF
INFORMATION
Table 3a on page 29 of Road Casualties Great Britain 2006
provides estimates of numbers of casualties in recent years
in accidents in which at least one driver had a BAC over 80.
Pending widespread use of new technology in screening breath
tests and in testing accident-involved drivers, and the possible
conduct of a fresh roadside survey of the BACs of drivers in general,
information about distributions of BACs of drivers killed and
drivers involved in injury accidents and estimated relationships
between accident risk and driver's BAC are drawn from Maycock's
1997 review. Confidence in the continued relevance of this information
is drawn from the general consistency between the unpublished
results of the 1998-99 roadside surveys (Tunbridge et al 2003)
and those of earlier surveys used by Maycock, and the broad stability
over time, and consistency with the data used by Maycock, of the
distribution of BACs of killed drivers reported in Table 3h of
Road Casualties Great Britain 2006 and corresponding tables
in earlier years. The author's earlier thinking about possible
effects of changes in the legal BAC limit upon numbers of accident-involved
drivers, which started from the corresponding discussion in the
consultation paper Combating drink driving: next steps, is
recorded elsewhere (Allsop 2005).
THE PROCESS
OF ESTIMATION
Table 3a of Road Casualties Great Britain 2006 shows
accidents in which at least one driver's BAC exceeds 80 accounting,
in round figures, for about 550 killed, 2,000 seriously injured
and 12,000 slightly injured per year. The annual number killed
has been broadly steady for a decade, during which the number
seriously injured has fallen by about one-third and the number
slightly injured has fluctuated between about 12,000 and 17,000.
Estimates of reductions made here will be based on there being
550 killed and 2,000 seriously injured per year in such accidents
with the present BAC limit and, as a conservative estimate, six
slightly injured for every one seriously injured. The recent downward
trend in serious injures in such accidents of about 4% per year
should be borne in mind when using the estimates of effect on
numbers injured. Estimated reductions corresponding to other current
annual numbers killed or injured in accidents in which at least
one driver's BAC exceeds 80 can be obtained by prorating the estimates
made here.
Assumptions which are described here as conservative are
so in the sense that they are likely to err on the side of underestimating
the casualty reductions to be expected from reducing the legal
limit. Other assumptions are made on the basis of being neutral
in this respect.
It is assumed that behaviour of those now driving well over
the existing limit of 80 is unlikely to be affected by lowering
the limit. It is therefore assumed that casualty reductions resulting
in the short term from reducing the limit will arise from reduction
in accident involvement of drivers below or only somewhat above
the current limit. (This is not to say that lowering the limit
will never affect the amount of driving at higher BACs, but recognises
that appreciable effects on such driving will come mainly through
longer term changes in the culture of drinking and driving, leading
to fewer people developing the habit of driving after heavy drinking).
The following four ranges of BAC are therefore considered: 80-110,
50-80, 20-50, and below 20 but not effectively zero (in that they
are so low that they are not necessarily related to drinking behaviour,
and in any case there is no reason in terms of accident risk for
seeking to change them).
For these four ranges, numbers killed per year in accidents
in which a driver's BAC lies in each range are estimated from
the total of 550 for which the BAC is over 80 using the distribution
of BACs in car drivers killed in 1990-94 in Table 10 of Maycock's
review. It is assumed that the distribution of BACs of drinking
drivers involved in fatal accidents is estimated by that of the
drinking drivers who were themselves killed. With access to the
data on which Table 3h of Road Casualties Great Britain 2006
and its predecessor tables are based, Maycock's Table 10 could
be updated and replaced in the analysis reported here.
Because the numbers of drivers in the BAC intervals 41-80
and 81-120 in Table 10 are so nearly equal, it is assumed that
drivers' BACs are uniformly distributed over the range 40-120.
It is further assumed for purposes of calculation that BACs that
are not effectively zero are uniformly distributed over the range
0-40 with the same density as between 40 and 120—a conservative
estimate because more drivers would be expected at these lower
BACs. It follows that the annual numbers of deaths in fatal accidents
with a drinking driver in each of the four BAC ranges being considered
are estimated to be as shown in Table 1.
The corresponding numbers seriously injured are estimated
similarly from the total of 2,000 for which the drinking driver's
BAC is over 80 using the distribution of breath alcohol concentrations
in drivers involved in injury accidents in Table 8 of Maycock's
review, after conversion to BACs. Whilst it is less clear than
for Table 10 (relating to drivers killed) how representative the
distribution in Table 8 is, it is the best available and was used
by Maycock as such. Its use in relation to serious injuries is
conservative in that it relates to all injury accidents, whereas
the proportions of drivers with higher BACs would be expected
to be greater among those involved in serious accidents only.
The resulting estimates of the annual numbers seriously injured
in accidents with a drinking driver in each of the four BAC ranges
being considered are estimated to be as shown in Table 1.
Table 1
ESTIMATED ANNUAL NUMBERS KILLED AND SERIOUSLY INJURED
IN ACCIDENTS WHERE A DRIVER HAS A BAC OF UP TO 110
| | |
| BAC of drinking driver | Number killed
| Number seriously injured |
| | |
| 80-110 | 65 | 557
|
| 50-80 | 65 | 557
|
| 20-50 | 65 | 557
|
| 0-20 but not effectively zero | 43
| 371 |
| | |
| |
|
If drivers' BACs are influenced by reducing the legal limit,
the effect on numbers killed or injured will be determined by
the resulting changes in risk of accident involvement. Maycock
estimated from his Tables 8 and 10 and the results of roadside
surveys in 1988 ad 1990 of BACs of drivers in Britain that the
risks of being killed and of involvement in an injury accident
at a BAC of b were proportional to exp(0.032b) and exp(0.021b)
respectively (after conversion from units of breath alcohol concentration).
In line with previous assumptions, these relationships will be
applied here to estimate changes in numbers killed and seriously
injured respectively.
In doing so, two sets of assumptions, respectively pessimistic
and optimistic in terms of casualty reduction, are made about
how drivers will change their behaviour in response to reduction
in the limit, in order to obtain lower and higher estimates of
casualty reduction. These assumptions, set out in Table 2, are
new to this process of estimation.
Table 2
ASSUMPTIONS ABOUT CHANGES IN DRIVER BEHAVIOUR IN RESPONSE
TO REDUCTIONS IN THE LEGAL BAC LIMIT
| | |
| Drivers' current BACs | Assumption about altered BACs
|
| Pessimistic | Optimistic
|
| | |
| Limit reduced to 50
| |
| 80-110 | Redistributed over 50-110 in the same order
| All reduced by 30 |
| 50-80 | Redistributed over 20-80 in the same order
| All reduced by 30 |
| 20-50 | Redistributed over 0-50 in the same order
| Redistributed over 0-20 in the same order |
| 0-20 | Unchanged | Unchanged
|
| Limit reduced to 20
| |
| 80-110 | Redistributed over 20-110 in the same order
| All reduced by 60 |
| 50-80 | Redistributed over 0-50 in the same order
| Those over 60 reduced by 60; those 50-60 reduced to 0
|
| 20-50 | Redistributed over 0-20 in the same order
| Those over 30 reduced by 30; those 20-30 reduced to 0
|
| 0-20 | Unchanged | Reduced to zero
|
| | |
| |
|
The effects of these assumptions, together with the previous
assumption that drivers are currently distributed uniformly over
each of the four ranges being considered, upon numbers of deaths
and seriously injured in accidents involving drivers currently
in these four ranges can be derived by integration in the form
of factors by which the numbers killed or seriously injured would
be reduced, as set out in the Annex.
ESTIMATED REDUCTIONS
IN CASUALTIES
Applying the factors derived in the Appendix to the numbers
in Table 1 yields the optimistic and pessimistic estimates of
reductions in annual numbers killed or seriously injured set out
in Table 3 for an existing situation in which the 550 are killed,
2,000 seriously injured and 12,000 slightly injured annually in
accidents in which at least one driver's BAC exceeds 80. The estimates
shown of the reduction in the number slightly injured are simply
conservative estimates obtained by multiplying by six the estimated
reductions in the number seriously injured.
Table 3
ESTIMATED ANNUAL REDUCTIONS IN CASUALTIES RESULTING FROM
REDUCING THE DRINK-DRIVE LIMIT TO 50 OR 20 MG/100ML
| | |
| |
| Drivers' current BACs | Assumption about altered BACs
| | | |
| Pessimistic
| Optimistic |
| Killed | Seriously injured
| Killed | Seriously injured
|
| | |
| |
| Limit reduced to 50
| | |
| 80-110 | 23 | 144
| 40 | 260 |
| 50-80 | 23 | 144
| 40 | 260 |
| 20-50 | 17 | 102
| 36 | 227 |
| 0-20 | 0 | 0
| 0 | 0 |
| Total | 63 | 390
| 116 | 747 |
| Slightly injured | | 2,340
| | 4,492 |
| Limit reduced to 20
| | |
| 80-110 | 36 | 240
| 55 | 399 |
| 50-80 | 47 | 315
| 55 | 393 |
| 20-50 | 36 | 227
| 39 | 249 |
| 0-20 | 0 | 0
| 9 | 68 |
| Total | 119 | 782
| 159 | 1,109 |
| Slightly injured | | 4,692
| | 6,654 |
| | |
| |
| |
| | |
Estimated reductions corresponding to other current annual
numbers killed or injured in such accidents can be obtained by
prorating the estimates in Table 3. For example, estimated reductions
of the year 2007, in which an estimated 410 were killed, 1,760
seriously injured and 11,850 slightly injured in such accidents
can be obtained by multiplying the reductions in numbers killed
by 410/550 = 0.75, the reductions in numbers seriously injured
by 1,760/2,000 = 0.88, and the reductions in numbers slightly
injured by 11,850/12,000 = 0.99.
REFERENCESAllsop R (2005) Some
reasons for lowering the legal drink-drive limit in Britain CTS
Working Paper 2005/1 London: Centre for Transport Studies,
University College London
Department for Transport (2007) Road Casualties Great Britain
2006 London: TSO
Department of the Environment, Transport and the Regions (1998)
Combating drink driving: next steps London
Maycock G (1997) Drinking and driving in Great Britain—a
review TRL Report 232 Crowthorne: Transport Research Laboratory
Tunbridge R J, M Keigan and F J James (2003) The distribution
of breath alcohol levels in drivers Unpublished Report PR SE/735/03
Crowthorne: TRL Ltd
Annex
ESTIMATED EFFECTS OF ASSUMED ALTERATIONS IN DRIVERS' BACS
UPON CASUALTY NUMBERS
The factors by which numbers of casualties will be multiplied
under the assumptions in Table 2 are each estimated by use of
one or two of the following four expressions, in which k = 0.032
for numbers killed and k = 0.021 for numbers seriously injured.
These expressions yield the values shown in Table A1, which
are independent of the numbers killed or injured in accidents
in which at least one driver's BAC exceeds 80.
Table A1
FACTORS BY WHICH NUMBERS OF KILLED OR SERIOUSLY INJURED
ARE ESTIMATED TO BE MULTIPLIED BY CHANGES IN BEHAVIOUR ASSUMED
IN TABLE 2
| | |
| |
| Drivers' current BACs | Assumption about altered BACs
| | | |
| Pessimistic
| Optimistic |
| Killed | Seriously injured
| Killed | Seriously injured
|
| | |
| |
| Limit reduced to 50
| | |
| 80-110 | .6428 | .7419
| .3829 | .5326 |
| 50-80 | .6428 | .7419
| .3829 | .5326 |
| 20-50 | .7386 | .8166
| .4512 | .5926 |
| 0-20 | 1.0 | 1.0
| 1.0 | 1.0 |
| Limit reduced to 20
| | |
| 80-110 | .4445 | .5685
| .1466 | .2837 |
| 50-80 | .2828 | .4349
| .1553 | .2943 |
| 20-50 | .4512 | .5926
| .4056 | .5526 |
| 0-20 | 1.0 | 1.0
| .7386 | .8165 |
| | |
| |
| |
| | |
| 
