Work Related Road Safety Literature Review

Work-related
road safety
A systematic review of the literature on the effectiveness of interventions
Report submitted to the IOSH Research Committee
Graham B Grayson BSc PhD and Shaun Helman BSc PhD

Transport Research Laboratory
Crowthorne House
Nine Mile Ride
Wokingham
Berkshire
RG40 3GA
UK
research report
11.3
IOSH, the Chartered body for All recipients of funding from
our Research and Development
health and safety professionals,
Fund are asked to compile a
is committed to evidence- comprehensive research report of
based practice in workplace their findings, which is subject to
health and safety. We maintain peer review.
a Research and Development
For more information on how to
Fund to support research and apply for grants from the Fund,
inspire innovation as part of visit www.iosh.co.uk/
our work as a ‘thought leader’ researchanddevelopmentfund,
or contact:
in health and safety.
Dr Luise Vassie
Executive Director – Policy
luise.vassie@iosh.co.uk
Mary Ogungbeje
Research and Technical Adviser
mary.ogungbeje@iosh.co.uk
Work-related
road safety
A systematic review of the literature on the effectiveness of interventions
Report submitted to the IOSH Research Committee
Graham B Grayson BSc PhD and Shaun Helman BSc PhD

Transport Research Laboratory
Crowthorne House
Nine Mile Ride
Wokingham
Berkshire
RG40 3GA
UK
All rights reserved. No part of this publication may be reproduced in any material form (including
photocopying or storing it in any medium by electronic or photographic means and whether or not
transiently or incidentally to some other use of this publication) without written permission of IOSH,
the copyright owner. Applications for written permission to reproduce any part of this publication
should be addressed to the publisher.
IOSH assumes no responsibility for the contents of this research report, in whole or in part, nor for
the interpretation or concepts advanced by the authors. The views expressed in this report are the
authors’ own, and do not necessarily reflect those of any employing or other organisation. All web
addresses are current at the time of going to press. The publisher takes no responsibility for
subsequent changes.
Warning: The doing of an unauthorised act in relation to a copyright work may result in both a civil
claim for damages and criminal prosecution.
This report is printed on chlorine-free, acid-free stock produced from woodpulp originating from
managed, sustainable plantations. The paper is recyclable and biodegradable.
© IOSH 2011
Printed in England by Paradigmprint (UK) Ltd
Published by IOSH
The Grange
Highfield Drive
Wigston
Leicestershire
LE18 1NN
UK
t +44 (0)116 257 3100
f +44 (0)116 257 3101
www.iosh.co.uk
Contents
Abstract 5
Executive summary 6
1 Background 9
2 Method 10
3 Literature review 11
4 Interventions to improve work-related road safety 14
5 Methodological issues 24
6 Summary 29
References 30
Work-related road safety: a systematic review of the literature on the effectiveness of interventions 5
Abstract
Road casualty statistics show that a large proportion of road casualties are accounted for by people
who are in some way driving for work, so interventions to improve work-related road safety (WRRS)
could have considerable potential. There are a number of different forms that interventions focused
on WRRS can take. Although there are many providers of such interventions, there is a lack of
understanding as to which interventions are most effective, and to what degree. A systematic review
of the literature on WRRS has therefore been carried out in order to make an evidence-based
appraisal of the effectiveness of WRRS interventions. Following extensive searches of the Transport
Research Abstracting & Cataloguing System, a total of 63 studies are discussed in this report,
including six earlier reviews dating from 1999 to 2011.
The review covered six main areas: driver training, group discussions, incentive schemes, publicity, in-
vehicle recorders, and organisational approaches. Although the study set out to provide evidence-
based advice to practitioners, this proved to be a surprisingly difficult task. Only four interventions
were found in studies of a scientifically acceptable standard that showed statistically meaningful
reductions in crash risk. Three were in the same investigation, and all were conducted more than a
decade ago. Possible reasons for this are discussed. These include changes in recent times in research
and procedures from single to multiple interventions, the trend to using attitudinal and behavioural
measures as dependent variables, and the increasing acceptance of case studies as evidence for positive
change. While it is accepted that there are commercial and practical issues in trying to persuade
industry to engage in evaluation studies, the main conclusion of the project is that there is a pressing
need for more and better-controlled evaluation work if a better understanding of WRRS issues is to
be achieved.
6 Grayson and Helman
Executive summary
Background
Road casualty statistics show that a large proportion of road casualties are accounted for by people
who are in some way driving for work. It has been estimated that between a quarter and a third of all
road traffic incidents involve someone who was at work at the time. Thus interventions to improve
work-related road safety (WRRS) could have considerable potential in terms of the absolute numbers
of lives that could be saved and injuries prevented.
There are a number of different forms that interventions focused on WRRS can take. Although there
are many providers of such interventions, there is a lack of understanding as to which interventions
are most effective, and to what degree. A systematic review of the literature on WRRS has therefore
been carried out that has attempted to answer the following questions:
• Is there a sufficient literature of the highest quality evaluations (ie randomised controlled trials
with sufficient sample sizes and using collisions or collision risk as an outcome variable) to make
a definitive statement regarding the efficacy of WRRS interventions overall?
• If not, what are the suggested levels of effectiveness of WRRS interventions using weaker study
designs?
• In either case, what can be said about the effectiveness of different subcategories of intervention
(such as training, incentives, and enforcement through technology)?
The main data source was the Transport Research Abstracting and Cataloguing System, which is the
main catalogue of publications held both in the TRL library and by other major transport research
organisations. It now comprises 260,000 items, and is the prime literature resource for transport
research. Following extensive searches, a total of 63 studies are discussed in this report, including six
earlier reviews dating from 1999 to 2011.
Literature on interventions
While organisations are now given a great deal of advice and encouragement to engage themselves
more actively in efforts to improve WRRS, very little has been published in the way of guidance on
how to achieve this aim. To assist in this process, this report sets out to make an evidence-based
appraisal of the effectiveness of WRRS interventions. The main findings from the review are set out
below.
Driver training
Traditionally, the first, and often the only, line of attack when attempting to reduce fleet driver
accidents has been through driver training. Driver training has proved to be one of the most
contentious topics in road safety for well over half a century. A recent review concluded that there
was no evidence that pre-licence training had any measurable effect on crash risk, and many of the
reports of the effectiveness of post-licence training are largely anecdotal. A review of post-licence
driver training carried out for the Cochrane Collaboration in 2003 concluded that there was no
evidence that post-licence driver education programmes were effective in preventing road traffic
injuries or crashes. In the fleet safety field, properly designed experiments are extremely rare; indeed,
only one appears to have been published in a peer-reviewed journal. This was the study published in
1996 using the Swedish telephone company Televerket.26 In this experiment, four interventions (driver
training, group discussions, campaigns and bonuses for accident-free driving) were compared with a
control group, and accident rates and costs were compared for a two-year period before and after the
interventions. The results for the driver training group showed a statistically significant reduction of
40 per cent in the accident rate after training, though the training procedures employed in the study
were far removed from those conventionally used by fleet trainers. While the Swedish study has been
widely cited in the literature, and is generally recognised as being highly innovative in terms of both
experimental design and in the content of training materials, there has been no documented attempt
to take this work forward.
Group discussions
One method that has aimed to raise safety consciousness and thereby bring about behavioural change
is the use of group discussions among employees. This method has not been widely used as a WRRS
intervention, but is one that would seem to offer promise. The results from the Swedish experiment
showed that group discussions were the most successful of the four interventions used in the study,
with a statistically significant reduction in accident rates of 56 per cent in the two years after the
measure was introduced. The group discussion approach has a sound theoretical basis, there is at
least one empirical study that attests to its effectiveness, and it should be attractive from an economic
point of view. It is unfortunate that properly controlled investigations have not been carried out to
take forward or even replicate the earlier Scandinavian research studies.
Incentives
The road traffic system operates largely on the principle that bad behaviour should be punished, but
good behaviour goes unrewarded. The corporate fleet environment is one where there is the
possibility of redressing this balance, and a number of studies have been carried out to look at the
effects of incentives or rewards on driver behaviour and accident rates. A bonus, or more properly an
incentive system, was one of the four measures studied in the Televerket experiment. There was a
significant 23 per cent reduction in the accident rate for the bonus group; there was also some
supportive evidence from a UK study that incentives had a beneficial effect. At an anecdotal level, it is
known that many organisations operate incentive and reward systems, but there is no published
evidence on the effectiveness of such schemes. In summary, it is once again unfortunate that there has
been so little follow-up of an intervention that has both theoretical and empirical support, is
relatively cheap and easy to implement, and should in principle lend itself to proper investigation.
Publicity
The Swedish study also included a ‘campaign’ group, which comprised five staff meetings during the
course of a year during which videos were shown and publicity material was distributed. The results
of the experiment indicated that this was the only one of the four test groups not to show a decrease
in accident rate relative to the control group; in fact, it showed a small but non-significant increase. A
review of US work covered a range of procedures, such as promise cards, performance feedback,
pledge cards and safety reminders. However, there is no hard evidence in the literature that the
interventions described above are effective in improving WRRS.
In-vehicle data recorders (IVDRs)

In-vehicle data recording systems have come into widespread use in recent times in occupational
settings. They were widely referred to originally as ‘black boxes’, partly because the early versions
were quite large, but mainly because the original use, as in aircraft, was to collect data in the period
immediately before a crash. While still used to some extent for accident investigation purposes,
advances in technology have led to an increasing role in driver monitoring. When it comes to hard
evidence, the picture is a familiar one of potential not being recognised. There is only one properly
controlled study in the literature to date that demonstrates a statistically significant reduction in
accident involvement in fleets as a result of installing IVDRs. This covered a number of fleets in
Belgium and the Netherlands, and drivers were made aware that the data recorded could be used
when accident responsibility was being assessed. Once again, it is necessary to use the word
‘unfortunate’ to describe a situation where an intervention that has great potential for research as
well as evaluation appears so infrequently in the literature. The opportunity to collect and use real-
time data in real-world situations should not be lightly dismissed.
Organisational approaches
All previous interventions have been concerned directly with change at the employee level in an
organisation. A major shift in research emphasis in recent times has been from how individuals
respond to interventions to how organisations plan interventions. A recent review lists a dozen or so
practices and procedures that have been recommended to organisations for the better management of
occupational road risk, but notes that many of these have not yet been scientifically evaluated. Of
those that had, it was found that designs were methodologically weak and lacked statistical rigour. It
would appear that the limited evaluation literature on the subject offers little in the way of scientific
support for interventions at the organisational level to improve WRRS.
Methodological issues
Having reviewed the literature on WRRS interventions, it has to be concluded that it is very
probable that there is not a single item in it that would be deemed acceptable for inclusion in a
Cochrane review. The basic reason is that randomised allocation – the central tenet of the
experimental method – is extremely difficult to achieve in the real world, and particularly in an
occupational or commercial setting. There are also numerous other problems that make it difficult
to achieve quality in evaluation studies in this area. While past experience has shown that these
problems are not insuperable at a practical level, in recent times problems of procedure have arisen
that act as barriers to evaluation.
One very obvious feature of changing practices in the last decade has been the shift from single to
multiple interventions, in which organisations are encouraged to adopt a ‘proactive multiple strategy
approach’. Such advice leads inevitably to methodological problems when it becomes impossible to
distinguish the unique impact of initiatives when they are implemented in combination with others.
Another trend has been the increasing amount of research in the last decade concerned with the
investigation of attitudinal and behavioural factors associated with crash involvement. The rationale
for this activity is that, if it can be shown that self-reporting measures are predictive of crash
involvement, then they would provide a richer and more informative outcome measure than the mere
counting of crash frequencies. There are a number of inherent problems with this approach, and it
has achieved only limited success to date. A realistic assessment should be that the prospect for
developing proxy measures for fleet accidents at present looks challenging.
A third development in the last decade is the rise of the case study. These typically describe success
stories within organisations, and have given considerable encouragement to those who wish to adopt
the ‘best practice’ approach. Practitioners readily point to case studies as clear examples of how the
organisational approach to fleet safety can achieve positive results. However, there are two issues that
should be borne in mind. First, only positive outcomes tend to be reported. There is anecdotal
evidence that some large organisations have implemented programmes without achieving any benefits.
If the ratio of successful to unsuccessful programmes is unknown, then the value of such interventions
must remain unclear. Second, there is the problem of generalisation. Case studies are a very good
method of showing what can be achieved by bringing about change, but they provide only limited
information on how these results can be transferred to other situations.
Experimentation to establish the effectiveness of remedial measures is extremely difficult. Twenty

years ago, this was seen as a challenge; good evaluation work was carried out, and possibilities for
further research were identified. These possibilities have never been followed up. Largely as a result of
pressure from policy makers and legislators, there has been a major change in emphasis where a
concern with policies and procedures has become the main priority for fleet operators. The final
barrier to evaluation is that it is no longer seen as being important – at least not in the rigorous
scientific sense. In the multifactor approach, changes to crash rates are now seen as only one of many
possible key performance indicators. One effect of these barriers to evaluation is the fact that not a
single properly controlled evaluation study of WRRS interventions has been carried out in the last
decade.
Conclusions
This study set out to provide evidence-based advice to practitioners about the effectiveness of
interventions to improve WRRS. This has proved to be a surprisingly difficult task. If one adopts the
criterion that an evaluation study should assess whether an intervention has brought about a
statistically reliable change in crash rates, then the results are meagre in the extreme. Only four
interventions meet this criterion, three were in the same investigation, and all were conducted more
than a decade ago.
It is important to bear in mind that the absence of strong evidence of effectiveness does not mean that
effectiveness has not been achieved. There have been a number of case studies in recent times that
have claimed impressive gains in WRRS through the implementation of large-scale programmes.
However, it is known that some unsuccessful case studies go unpublished, and this, combined with
the fact that such large-scale programmes tend to be multifaceted, means that such success stories can
contribute little in the way of understanding about the effectiveness of different intervention
components. Without a return to ‘classical’ evaluation, work to improve WRRS can only be on an ad
hoc and less than efficient basis. There are of course commercial and practical issues that remain to
be overcome when trying to persuade industry to engage in evaluation studies. Companies may not
wish to invest time and money in evaluation studies if they perceive that holistic approaches (even if
only supported by case study data) are sufficient for their needs. In addition, companies may be
reluctant to have their accident data made public. Nonetheless, the fact remains that there is a
pressing need to use controlled evaluation studies to assess which WRRS interventions work, by how
much, and through which causal mechanisms.
1 Background
1.1 The problem

Road traffic collisions are a serious public health burden. According to the World Health
Organization,1 they are predicted to be the fifth-highest cause of death by 2030, as compared to being
only the ninth-highest cause in 2004.
The statistics show that a large proportion of road casualties are accounted for by people who are in
some way driving for work. For example, figures provided by the Department for Transport show
that in Britain in 2009, 18 per cent of all drivers and riders aged 15 or over who were involved in a
collision where someone was injured were ‘driving for work’ at the time. An earlier estimate by the
Work-related Road Safety Task Group2 was that between a quarter and a third of all road traffic
incidents involved someone who was at work at the time. Thus interventions that improve work-
related road safety (WRRS) could have considerable potential in terms of the absolute number of lives
that could be saved and injuries prevented.
1.2 Why this review is needed

There are a number of different forms that interventions focused on work-related road safety can
take. Although there are many providers of such interventions, there is a lack of understanding as to
which types of interventions are most effective, and to what degree. In other fields dealing with health
outcomes, systematic reviews of empirical evidence such as those carried out by the Cochrane
Collaboration (see www.cochrane.org) have been accepted as being the best way to establish the level
of support for any given intervention or treatment. Taking this as a model, a systematic review of the
literature on WRRS has been carried out that has attempted to answer the following questions:
• Is there a sufficient literature of the highest quality evaluations (ie randomised control trials with
sufficient sample sizes and using collisions or collision risk as an outcome variable) to make a
definitive statement regarding the efficacy of WRRS interventions overall?
• If not, what are the suggested levels of effectiveness of WRRS interventions using weaker study
designs?
• In either case, what can be said about the effectiveness of different subcategories of intervention
(such as training, incentives, and enforcement through technology)?
2 Method
2.1 Inclusion criteria

A preliminary search of the published literature led the authors to believe that there would be very
little that would meet the methodological standards required by standard systematic reviews, such as
random allocation to treatment or control groups, and proper statistical analysis of effect size. With
this in mind, weaker designs were included as long as they were still evaluations of effectiveness on
the outcome measures of collisions, behaviours or attitudes. Recently there has been criticism (eg
Wåhlberg et al.3) of the validity of self-reported collisions because of under- and over-reporting
(arising from factors such as memory bias and social desirability) and their generally lower than
desirable convergence with recorded collisions (although see Boufous et al.4). This debate is not new,
and it is also not settled (see Arthur et al.5 for a useful overview). On balance, there would appear to
be a reasonable consensus in the literature that both types of data can be used as outcome measures,
even if they both have their limitations (see Maycock et al.,6 Arthur et al.,5 Boufous et al.4). Therefore
studies using recorded collisions and those using self-reported collisions were both considered for
inclusion in the review.
In addition, published studies that were not strictly outcome evaluations of effectiveness, but instead
used a weaker case-study design, were included. However, due to the inherent limitations of the case-
study design (see Coolican7), it was decided that detailed descriptions of case studies – for example
the precise package of measures used – would be avoided.
In addition to presenting a review of WRRS interventions as described above, the report includes a
commentary on how the WRRS field has developed over the last decade or so.
2.2 Search methods and terms

The following search terms were used:
(Work OR Occupational) AND (Road OR Driving) AND (Safety OR Risk) Fleet AND Safety
“Work related road safety”
These search terms were entered into the TRL Knowledge Base. This comprises a number of
databases including the Transport Research Abstracting and Cataloguing System (TRACS), which is
the main catalogue of publications held both in the TRL library and by other major transport
research organisations. It contains bibliographic references and abstracts of English and foreign
language articles from journals, books and research reports. It is the English-language version of the
worldwide International Transport Research Documentation (ITRD) database and contains abstracts
from publications in the USA, Australia, Scandinavia, the Netherlands and Canada, in addition to UK
material. The database has been updated daily since 1972 and now comprises 260,000 items. This is
the prime literature resource for transport research.
In addition, various searches using Google, Google Scholar, and known lists of ‘grey literature’ (for
example, SIGLE – System for Information on Grey Literature in Europe) were carried out, and work
referenced by previous reviews and by included studies were followed up for additional material.
2.3 Data collection

The studies identified from the searches were examined by the two authors for eligibility based on the
inclusion criteria. The full text of all relevant articles was obtained and their findings were
synthesised by the lead author, reviewed by the second author, and finally checked by a third reviewer
not on the report team.
3 Literature review
3.1 Introduction
In recent years, the safety of people who drive for work has received increasing attention from policy
makers and road safety professionals. The domain of WRRS is generally taken to be that of any
journey undertaken in the context of work (Lang et al.8). Two points should be borne in mind. First,
commuting to the normal workplace is not generally treated as work-related driving. A recent review
by Murray9 found that this is the case in most developed countries, although there are exceptions,
including Finland and some Australian states. Second, journeys carried out in privately owned
vehicles for work purposes are included as being work-related. These vehicles, sometimes referred to
as ‘grey fleets’, have long been recognised as posing special problems for safety management.
The management of WRRS is important because statistics show that a sizeable proportion of road
casualties are accounted for by people driving for work. In Britain, journey purpose has since 2005
been included in the accident reporting process. Figures provided by Department for Transport show
that in Britain in 2005, 15 per cent of all drivers or riders aged over 15 and involved in a collision
where someone was injured were ‘driving for work’ at the time. This figure is almost certainly an
underestimate, given the recent introduction of the new variable into the reporting process, and the
possible reluctance of some light vehicle drivers to report at-work collisions (Lang et al.8). In 2009
(the most recent year for which data are available), the figure was 18 per cent.
In the context of occupational safety, the contribution of work-related driving is even more striking.
Figures from the EU (excluding the UK) show that over one-third of fatal accidents at work occur on
the roads (ETSC10), while Murray9 reports a similar proportion in the United States. Occupational
safety statistics in the UK are complex, but Lang et al.8 cite data indicating that work-related driving
fatalities far outstrip the number of fatal injuries that occur in the ‘traditional workplace’.
The efficient management of WRRS therefore has the potential to make a significant contribution to
the reduction of death and injury on the roads. To assist in this process, this report sets out to make
an evidence-based appraisal of the effectiveness of WRRS interventions. It begins by looking at how
work-related driving was seen in the past, and the changes that have taken place in the last two
decades. It then goes on to examine in detail the interventions that have been employed to improve
WRRS, and to make a critical assessment of their effectiveness. Methodological issues are then
discussed before conclusions are presented.
3.2 The past (1970s to mid-1990s)

The term ‘work-related road safety’ has only appeared relatively recently in the road safety literature.
In the latter part of the last century, the view was generally held in the UK that heavy goods vehicles
were not a major road safety problem. The freight industry was regulated and monitored, its vehicles
were subject to frequent inspection, and its drivers were required to undergo training and testing that
was more rigorous than that required of the rest of the driving population. Of more concern to
researchers and policy makers was the rapid growth in car ownership during this period, with an
expanding driver population that faced only relatively undemanding barriers to entry, and whose
members only became of interest to the authorities if they transgressed in a serious way.
Britain was (and still is) unlike most other European countries in that a significant proportion of the
total vehicle fleet is made up of cars that are owned or financed by commercial organisations. The
drivers of company cars made up an important subgroup of the driving population, and concerns
were expressed by policy-makers that they made a disproportionate contribution to casualty statistics.
In response to this, there was a flurry of research activity in Britain in the 1990s. The first major
study to be reported was by Lynn & Lockwood,11 who surveyed drivers who regularly drove a
company-owned or financed car. Using multivariate analysis, they found that such drivers had an
accident liability that was substantially greater than that of ‘ordinary’ drivers (a multivariate
approach allows for the complex interactions between variables under consideration). These results
were re-examined shortly afterwards by Downs et al.12 in the context of a review of fleet driver safety.
They concluded that:
• fleet car drivers do have an elevated accident liability even when mileage and demographics are
taken into account
• this effect can be quantified
• the size of the effect depends on the definition used.
Thus, a broad definition that included drivers of ‘perk’ cars gave a 29 per cent increase in accident
liability, while restricting the analysis to those who drove regularly for work purposes increased the
figure to between 40 and 50 per cent. Downs et al.12 also pointed out the diverse nature of the topic.
They noted that fleets varied enormously in both size and function, and their statement that ‘it could
be said that there is no such thing as a typical fleet driver’ has been frequently quoted.
Other researchers during this period also raised the issue of diversity. For example, Dimmer & Parker13
drew attention to the fact that ‘company car drivers include a range of road users from senior
executives provided with a second car as a perk of the job, through those who drive non-liveried
company-owned vehicles both for work and non-work purposes, to those employed to drive fleet cars,
vans or other specialist vehicles’. A study by Chapman et al.14 looked at drivers within a single
organisation and identified five distinct subgroups. There were differences among these groups in terms
of mileage, journey purpose, accident involvement and accident type. These findings point to the fact
that any elevated accident liability is unlikely to apply equally to all drivers in an organisation, but will
be dependent on the type of vehicle and the purpose of their journeys. This disparate nature of fleet
vehicles, fleet drivers and fleet activity has plagued research to the present day.
Grayson,15 in reviewing the early studies in this field, noted that an elevated accident liability, or ‘fleet
driver effect’, had been identified but not explained. Of all the possible explanations that had been
advanced in the past, only exposure in terms of the mileages travelled by fleet drivers was excluded
by the statistical modelling; other variables were examined, but the results were inconclusive. Thus
the 1990s ended without any clear research evidence that could guide the design of interventions to
improve fleet safety.
3.3 The present (mid-1990s to present day)

The last 15 years have been a time of change in the transport field. Relevant to this study are changes
in the vehicle fleet, travel patterns, employment practice, legislation and policy, and research
orientation.
One important change to the vehicle fleet in the UK has been the marked increase in the number of
vans on the roads. A study by Lang & Rehm16 found that the van population in the UK had grown
by a third over the preceding decade, and van traffic had increased by 40 per cent. The significance of
this finding is that light goods vehicles and their drivers are far less regulated and monitored than is
the case with the heavy goods sector. As far as changes in travel patterns are concerned, the authors
point out that ‘one of the reasons for the increasing numbers of vans on UK roads in recent years is a
significant growth in home shopping, as the majority of home shopping goods are delivered by vans’.
This, of course, is to a large extent a result of the growth in internet shopping. Another factor that
has implications for the management of safety is a change in employment practice. Outsourcing has
become more common, and deliveries are increasingly made by subcontractors or even self-employed
drivers.
As far as legislation and policy are concerned, the interface between occupational safety and
traditional road safety has not always been an easy one. As long ago as 1974, the Health and Safety
at Work etc Act established (where ‘reasonably practicable’):
• a requirement for safe plants and their maintenance

• a requirement for employers to provide information, instruction, training and supervision as
necessary to safeguard workers
• a requirement for employers to conduct their undertaking in such a way that the health and safety
of persons not in employment are not exposed to risks.
These requirements were subsequently strengthened under the Management of Health and Safety at
Work Regulations 1999, when employers became responsible for making suitable assessment of the
risks to their employees and to other persons, as well as for providing their employees with
information on risks to their health and safety and on preventive and protective measures. While
these regulations should in principle cover work-related driving, in practice the protection of workers
and the public in the context of road traffic law has been carried out by the police and the courts.
This is because the general policy has been not to invoke health and safety legislation when there is
existing law that protects workers and the public. This situation was felt by many to be
unsatisfactory, partly because available legislation was seen as not always being properly enforced,
and also because it led to a situation where the full extent of the problem could not be ascertained
because the data were not being collected systematically.
In an attempt to remedy this situation, a Work-related Road Safety Task Group was convened. Its
report2 included an estimate that between a quarter and a third of all road traffic incidents involved
someone who was at work at the time. Two years later, the Health and Safety Executive and the
Department for Transport issued joint guidance on work-related road safety,17 in which it was made
explicit for the first time that a vehicle being driven on the road for work purposes is part of the
workplace and therefore subject to health and safety regulation.
This guidance note had wide-ranging effects. For a start, the term ‘WRRS’ became firmly established
in the road safety literature (though it had been used in occupational safety for some time: see eg
Bibbings18). There was also a noticeable shift in the type of research activity, in that the focus of
interest moved from the driver to the organisation. Finally, WRRS has become big business.
Organisations are now bombarded with advice and exhorted to adopt a variety of initiatives,
frequently on the basis that by doing so they will minimise the threat of enforcement action or
litigation. The evidence supporting these initiatives will be examined in the following section.
4 Interventions to improve work-related road safety
4.1 Introduction
While organisations are now given a great deal of advice and encouragement to engage themselves
more actively in efforts to improve WRRS, very little has been published in the way of guidance on
how to achieve this aim. Since the Downs et al.12 review referred to in the previous section, there have
been only five publications that have provided comprehensive reviews of WRRS interventions: those
by Haworth et al.,19 Murray et al.,20 Murray9 and, most recently, by Banks et al.21 and Newnam &
Watson.22 Interestingly, all five emanate from Australia, reflecting a very active research and policy
interest there, or, as Murray9 put it, ‘a vibrant occupational road safety theme’. It should be noted
that of these five studies, only the Banks et al.21 review can be considered a properly critical review of
the literature, the others being largely descriptive in style.
The present study draws upon these six reviews, as well as on a large number of individual studies
identified in the search of the literature outlined in Section 3.2.
4.2 Driver training

Traditionally, the first, and often the only, line of attack when attempting to reduce fleet driver
accidents has been through driver training. As Downs et al.12 pointed out:
…fleets should in principle be well-placed to provide evidence on this issue given their high
accident rates, the degree of control they have over their drivers, and their potential to collect
reliable accident data.
As will be seen, practice has proved to be very different.
Driver training has proved to be one of the most contentious topics in road safety for well over half a
century. Advocates of the technique hold that it is virtually self-evident that improving driving skills
will have a safety benefit, and this view is widely accepted by the general public. As far as new
drivers are concerned, the evidence to support this contention has proved to be elusive. Numerous
reviews have been carried out on this topic, and the general consensus is that conventional driver
training has little or no effect on road safety in terms of a reduction in risk for new drivers. The most
recent review by Helman et al.23 concluded that ‘according to the evidence, it [driver training] has no
measurable effect on collision risk, and its continued use should therefore be set against much lower
expectations of what it can contribute directly to the safety of new drivers’.
Post-licence training, it is often argued, is different. It would appear that many would agree, given the
size of the training industry and the importance attached to training in the advice given to commercial
organisations. Unfortunately, many of the reports of the effectiveness of post-licence training are largely
anecdotal. For example, Haworth et al.19 describe six programmes carried out in the US in the preceding
decade, none of which provided adequate evaluation data. They also cite a report from the Herz
Corporation in 1995 that claimed to have achieved a 35 per cent reduction in crashes through using
defensive driving courses. The effectiveness of the study design was unclear, long-term effects were not
studied, and the results are at variance with much of the rest of the literature. Defensive driving courses
have been employed in the US for many years based on the seemingly plausible notion that teaching
drivers how to avoid accidents will prevent accidents. However, there are theoretical grounds for
questioning this, and extensive research has failed to demonstrate that such courses have any consistent
effect in reducing accident rates (the classic evaluation paper on this topic is by Lund & Williams24).
Post-licence driver training has also been the subject of a Cochrane systematic review; such reviews
are widely regarded as representing the ‘gold standard’ for methodological rigour. The review of post-
licence driver training carried out for the Cochrane Collaboration by Ker et al.25 considered only
randomised control trials that had been reported up to 2002, and included 24 studies, though noting
that ‘the methodological quality of included trials was generally poor’. Their conclusion was that:
This systematic review of randomised control trials provides no evidence that post-licence driver
education programmes are effective in preventing road traffic injuries or crashes. Because of the
large numbers of randomised participants included in the meta-analysis (close to 300,000) we can
exclude, with reasonable precision, the possibility of even modest benefits.
Such precision is in marked contrast to the claims sometimes made by training providers.
There is a certain irony associated with this review in that it was sponsored by a large multinational
company that had invested in driver education programmes and wished to establish whether this
policy was effective. The authors’ view was that ‘whilst we cannot claim that our results show that
this policy is ineffective, we would argue that … unrealistic expectations about the effectiveness of
driver education must be avoided.’
The studies in the Cochrane review were predominantly longitudinal in design. By contrast, the
survey of company car drivers by Lynn & Lockwood11 referred to earlier was cross-sectional in
design. They included driver training as a variable in their multivariate analysis of self-reported
accidents, and found that trained drivers had an accident liability that was 8 per cent lower than that
of untrained ones, though the difference was not statistically significant.
In the fleet safety field, properly designed experiments are extremely rare; indeed, only one appears to
have been published in a peer-reviewed journal. This was the study of Televerket, a Swedish telephone
company, by Gregersen et al.,26 which Murray et al.20 have described as being ‘probably the most
quoted – and misquoted – fleet safety study undertaken anywhere to date’. In this experiment, four
interventions (driver training, group discussions, campaigns and bonuses for accident-free driving)
were compared with a control group, and accident rates and costs were compared for a two-year
period before and after the interventions. The results for the driver training group showed a
statistically significant reduction of 40 per cent in accident rate after training. This is convincing
evidence, but it should be borne in mind that the training procedures used in the study were far
removed from those conventionally used by fleet trainers. There were three components to the one-
day training programme in the Swedish study: low speed manoeuvring, skid training and commentary
driving. These are described in more detail in a later paper by Gregersen,27 who comments that:
...the large accident reduction was unexpected. The most probable explanation is the purpose and
content of the training. Specifically, the aim was not primarily to increase the driver’s skills in
manoeuvring the car, but to create insight about risks in traffic and about the driver’s own
limitations’.
He went on to express the hope that the study would ‘serve as an inspiration to develop such training
strategies further in combination with making scientifically correct evaluations’. Televerket was
unable to provide a venue for this. Not long after the study, the company was privatised, underwent a
major reorganisation, and the opportunity for any longer-term evaluation was lost.
One study that did attempt to follow up this innovative approach to driver training was carried out
in Finland by Salminen.28 Again, a one-day training course was employed, this time in ‘anticipatory
driving’, which focused on the problems of driving on snow and ice and in darkness. The main
outcome measure used in the study was an audit of work-related traffic safety in the company. There
was a statistically significant improvement in audit scores after the training intervention, and on this
basis the author claimed that the anticipatory driving course was successful. Further support was
claimed on the basis that drivers reported using what they had learned while on their holidays. Less
supportive of this position was the fact that traffic accidents increased after the intervention, albeit
based on very small numbers. The results must therefore be considered at best equivocal.
Less equivocal are the results from a recent study by Darby et al.,29 who carried out a retrospective
analysis of data from the records of a large organisation – a telephone company, as in the earlier
Swedish study. Their analysis showed a statistically significant 64 per cent reduction in claims rates
for drivers after undergoing driver training (no information was given on the nature of the training).
The analysis was rigorous and the results impressive, but the authors are rightly cautious in
presenting their conclusions given the methodological questions that can be raised about the study.
First, the drivers that were selected for training had claims rates that were well above the company
average, which raises the possibility of regression to the mean effects. Second, the design of the study
did not incorporate a control group. Third, and perhaps most important, driver training was only one
of a wide range of WRRS interventions introduced during the course of the study (according to the
authors, there had been no fewer than 124 safety interventions in the preceding five years).
A decade ago, Downs et al.12 reached the conclusion that ‘there is no evidence in the literature in the
form of scientifically controlled studies that conventional fleet driver training as practised in the UK is
effective in reducing accident rates’. A quote from a recent European report30 suggests that little has
changed: ‘There is no scientific evidence in the literature in the form of scientific controlled studies
that conventional fleet driver training as practised in the UK is effective in reducing crashes.’ The key
word in these two very similar quotes is ‘conventional’. As mentioned earlier, the Gregersen et al.26
study has been widely cited in the literature, and is generally recognised as being highly innovative in
terms of both experimental design and in the content of training materials. Despite this, there has
been no documented attempt to take this work forward, and to build on either the ‘insight’ approach
advocated by Gregersen, or to incorporate the proven benefits of training in the cognitive components
of driving, such as hazard perception (Helman et al.23).
4.3 Group discussions

One method that has aimed to raise safety consciousness and thereby bring about behavioural change
is the use of group discussions among employees. This method has not been widely used as a WRRS
intervention, but is one that would seem to offer promise. The theory that group decisions can be a
force for behavioural change dates back more than 60 years (Lewin31), but was only put into practice
in the occupational safety field for the first time in the Televerket study of Gregersen et al.,26 when it
was one of the four interventions in their comparative study. This component of their investigation
drew directly upon work carried out in Japan by Misumi,32 which in turn was based on the theories
of Lewin. Misumi described a study in which bus drivers participated in two discussions about work
problems, and then made decisions about a safer driving style. Accidents were reported as declining
sharply as a result. However, it should be noted that the drivers were included in the study because of
their prior accident involvement, so the possibility of regression to the mean should be borne in mind.
The Swedish study largely followed the procedures used by Misumi:
• a 60-minute ‘warm-up’ period of free discussion

• a 40-minute discussion to identify problems at the workplace
• a 20-minute meeting to discuss results of the previous stage
• small group discussions of the problems and how to resolve them
• a 60-minute plenary session to plan and commit to future action.
The results from the Swedish experiment showed that group discussions were the most successful of
the four interventions used in the study, with a statistically significant reduction in accident rates of
56 per cent in the two years after the measure was introduced. This conclusion has often been
reported as demonstrating the need for wider employee participation in WRRS. It is, however, worth
noting Gregersen’s comments in his later paper,27 in which he explains:
The group discussion intervention was a complex one … it is not possible to draw any definite
conclusions about the reason for the effect without further research to disentangle the effects of
the various factors.
Group discussions were also used in the Finnish study by Salminen.28 The procedures used were based
closely on the earlier Swedish investigation, the main difference being that they were conducted over a
six-month period. When comparing the numbers of accidents that occurred in the three years before
and after the intervention, the quoted result was a striking reduction of 72 per cent. Less often
quoted is the fact that this change was based on absolute numbers of 18 before and 5 after. The
author rightly points out that ‘the number of accidents was rather small … it is important that the
results of this study are in line with previous research in Japan and Sweden’.
In a different setting, group awareness sessions formed part of a long series of studies carried out in
the US by Ludwig & Geller33 that aimed to use behaviourist theories to bring about changes in safety
behaviour. Given the theoretical background, it is perhaps not surprising that the outcome measures
used were based on behavioural rather than accident measures; more importantly, the group
discussions formed part of a wider programme, and thus could not be evaluated separately.
In summary, the group discussion approach has a sound theoretical basis, there is at least one
empirical study that attests to its effectiveness, and it should be attractive from an economic point of
view. The suggestion by Gregersen27 that the driver training and group discussion approach might
profitably be combined has never been reported in the published literature (although it is said to have
been taken up by some organisations), and it is unfortunate that properly controlled investigations
have not been carried out to take forward or even replicate the earlier Scandinavian research studies.
4.4 Incentives
According to classical psychological theory, positive reinforcement is more effective than negative
reinforcement. Thus rats, pigeons and humans all respond better to reward than to punishment.
However, the road traffic system operates largely on the principle that bad behaviour should be
punished, but good behaviour goes unrewarded. The corporate fleet environment is one where there
is the possibility of redressing this balance, and a number of studies have been carried out to look at
the effects of incentives or rewards on driver behaviour and accident rates.
A bonus, or more properly an incentive system, was one of the four measures studied in the
Gregersen et al.26 Swedish experiment. There was a significant 23 per cent reduction in accident rate
for the bonus group. The authors felt that this was relatively modest compared with the other two
successful measures (driver training and group discussions), and questioned whether the level of
reward involved was too small given the constraints of Swedish taxation rules. Some evidence on this
point comes from the survey of company car drivers in the UK by Lynn & Lockwood.11 They found
that only a few of the companies in their sample provided any rewards or incentives for accident-free
driving, but that drivers in such schemes had fewer accidents than those that were not. Though only
marginally significant (p = 0.09), the effect was a large one, with a 21 per cent difference in accident
liability. Interestingly, the sums involved were generally small, and some rewards were non-monetary.
The possibility of cultural differences between Swedish and British drivers in this respect has not been
investigated further.
Wilde34,35 has long been an advocate of incentive schemes, and has argued that they should be used
more widely, both in occupational settings and to improve driver safety. Schneider36 describes a long-
term (30-year) programme in which professional drivers in a large German company were offered
incentives for accident-free driving. The result was a marked reduction in accident rates and costs
over the period. However, the author notes that this was only one of a variety of measures used by
the company.
The proceedings of an OECD symposium on enforcement and rewards (Koornstra & Christensen37)
contain several papers discussing incentive schemes from both theoretical and practical viewpoints,
but little has appeared in the literature since that time. Newnam et al.38 reported that financial
incentives based on insurance premiums had no effect on changing the attitudes of Australian fleet
safety managers, though the value of this outcome measure is difficult to assess given the lack of a
quantified link between managers’ attitudes and drivers’ accidents. Banks et al.21 describe a survey of
the US trucking industry that found that the safest firms used a range of driver reinforcement schemes
to encourage safer driving. The authors point out that failing to investigate less safe firms at the same
time considerably weakens this conclusion, and it is hard to see how it could be adduced as
‘evidence’.
At an anecdotal level, it is known that many organisations operate incentive and reward systems, but
there is no published evidence on the effectiveness of such schemes. Similarly, there is no evidence on
disincentive schemes, although these are known to be used – usually as part of a package of
measures. In summary, it is once again unfortunate that there has been so little follow-up of an
intervention that has both theoretical and empirical support, is relatively cheap and easy to
implement, and should in principle lend itself to proper investigation.
4.5 Publicity
The Swedish Televerket study by Gregersen et al.26 included what they termed a ‘campaign’ group.
This was not a conventional publicity campaign, in that it comprised five staff meetings during the
course of a year. In these meetings, seasonal problems for driving were discussed, videos were shown,
and publicity material was distributed. The results of the experiment indicated that this was the only
one of the four test groups not to show a decrease in accident rate; in fact, it showed a small but non-
significant increase. To confuse the picture further, the authors noted a small reduction in accident
costs, as well as a high drop-out rate for participation in the publicity meetings.
Community road safety campaigns in an organisational setting were also reported in the review by
Ludvig & Geller39 of behavioural change interventions over a 10-year period. Their review covered a
range of procedures, such as promise cards, performance feedback, pledge cards, and safety
reminders. The contribution of this research is limited by its focus only on behavioural outcome
measures, and by the absence of any tests of statistical significance in the results. Further, Murray et
al.20 rightly raise concerns that the 10-year study of pizza delivery drivers drew no attention to the use
of young staff in a productivity system that could be seen to encourage unsafe driving practices.
In summary, there is no hard evidence in the literature that the interventions described above are
effective in improving WRRS.
4.6 In-vehicle data recorders (IVDRs)

It was mentioned earlier in the review that the freight industry has tended to be more regulated and
monitored than other areas of work-related driving. One such regulation was the compulsory fitting
of tachographs, primarily for the purpose of monitoring driver hours. While an early American study
(Larson et al.40) claimed to have demonstrated a reduction in police vehicle accidents following the
installation of tachographs, the investigation was methodologically weak (poor control, inadequate
statistics and so on) and the conclusions no more than suggestive. Of greater relevance to the present
review are the newer IVDR systems that have come into widespread use in recent times in
occupational settings. They were widely referred to originally as ‘black boxes’ partly because the early
versions were quite large, but mainly because the original use, as in aircraft, was to collect data in the
period immediately before a crash. While still used to some extent for accident investigation purposes,
advances in technology have led to an increasing role in driver monitoring. The special characteristics
of IVDRs mean that they can objectively and accurately record the actual behaviour of identified
drivers in a naturalistic setting, which should make them ideally suited to learning more about the
actual, as opposed to the self-reported, behaviour of such drivers. There is the further possibility that
such devices could be used in the training of fleet drivers and to assess the effects of training
interventions by continued monitoring of drivers.
An examination of the evaluations that have been made of this potentially valuable intervention type
provides the by now familiar picture of enthusiastic promotion and limited hard evidence. A
European report41 talks of the possibilities of ‘empowering drivers to manage their own safety by
giving instantaneous in-vehicle feedback’, but then goes on to observe that ‘not all companies want to
reveal data as their technology gives them a comparative advantage over their competitors’. Concerns
over data sensitivity (both on the part of technology providers and users) may go some way to
explaining the lack of published evaluations using such promising technologies.
As far as evaluation is concerned, the report describes a ‘well-documented field test’ in which the
Berlin police department equipped all its patrol vehicles with IVDRs and claimed a 20 per cent
reduction in accidents in the following year. However, there was no control group included. The
report then goes on to state that:
...it became clear how important human leadership is in connection with the event recorders. Only
if staff or its representative body is involved early on and an awareness of joint responsibility for
the operational success can be achieved … can distrust and tensions be avoided. However, positive
impacts tend to fade out if monitoring and management efforts are not maintained at a high level.
This cautiousness is mirrored in an unpublished UK review on the topic, which noted numerous
limitations in the available information:
• several reviewed studies looked at only small samples, and over short periods of data collection
• the context of the installation, feedback and any external reference to the IVDR system may exert
its own distinct influences on driver behaviour
• some IVDR systems claimed to measure risk without demonstrating predictive validity
• several studies may have been subject to a regression to the mean effect when IVDR systems were fitted
in response to a high accident rate, and may thus have over-estimated the safety effects of IVDRs.
The European report also refers to positive results from programmes in the Netherlands and
Switzerland, but the information provided is anecdotal at best. When it comes to hard evidence, the
picture is a familiar one of potential not being recognised. A study by Toledo & Lotan42 showed that
‘risky’ manoeuvres among fleet drivers were reduced following the use of IVDR systems, but that
effects on behaviour tended to diminish over time. There is only one properly controlled study in the
literature to date (Wouters & Bos43) that demonstrates a statistically significant reduction in accident
involvement in fleets as a result of installing IVDRs. This was an output from an EU project
(SAMOVAR44), in which accident data recorders were installed in a number of fleets in Belgium and
the Netherlands, and drivers were made aware that the data recorded could be used when accident
responsibility was being assessed.
It should be noted that this study used what might be termed the first generation of IVDRs. Despite
technological advances in the capabilities of such devices, there appears to have been little in the way
of serious attempts to employ IVDRs in an evaluation context. One exception is the study by Toledo
et al.45 While essentially a demonstration project, the investigation included an analysis of crash rates
before and after the installation of IVDRs and found a statistically significant reduction in crash rates
overall (but not in at-fault crashes). The authors commendably point to the relatively short periods of
time involved, and to the fact that the results are somewhat at variance with their earlier work as far
as the stability of the effect is concerned. They suggest that ‘further research is needed to better
understand the temporal and long term impact of the installation and to develop feedback
management schemes … to maximise its impact’.
Once again, it is necessary to use the word ‘unfortunate’ to describe a situation where an intervention
that has great potential for research as well as evaluation appears so infrequently in the literature. The
opportunity to collect and use real time data in real world situations should not be lightly dismissed.
4.7 Organisational approaches

All previous interventions have been concerned directly with change at the employee level in an
organisation. As indicated earlier, a major shift in research emphasis in recent times has been from
how individuals respond to interventions to how organisations plan interventions. With the increasing
recognition of employer responsibility and duty of care, there has been a move to what has been
termed ‘work-related driver safety management approaches’ (Haworth et al.19). Associated with this
has been a growth in what might be termed ‘management-speak’, so that the literature is now replete
with best practice guidelines, safety policies and procedures, benchmarking, safety climate measures,
theoretical frameworks, matrices, case studies, human behaviour interfaces, proactive
multidimensional inputs, and even the ‘holistic’ approach.
‘Best practice’ is a case in point. Logically, the use of the word ‘best’ should imply that sets of
practices have been compared, their worth assessed in some way, and that one set has emerged as
superior to the others. In reality, these procedures are assembled largely on an a priori basis; ‘good
practice’ would be a more appropriate term, but presumably has less commercial appeal.
Furthermore, the use of the word ‘best’ runs the risk of conveying the implicit message that nothing
more needs to be done once such practices have been put into place.
What is missing in this plethora of initiatives is a commitment to evaluation. The Banks et al.21 review
lists a dozen or so practices and procedures that have been recommended to organisations for the
better management of occupational road risk. They then add a cautionary note: ‘as many of these
recommended initiatives have not yet been scientifically evaluated, enthusiastic endorsement of these
guidelines is cautioned’. Using their selection criteria, they found only four organisational
interventions that have been subjected to any form of evaluation: policy development, driver
selection, web-based tools and remuneration policy. Banks et al.21 found that evaluations of policy
development were weak in design and lacked statistical rigour, while the only study on driver
selection (from the US) lacked adequate controls. Similarly, while the web-based risk management
tool was found in a case study to be associated with a reduction in crashes (White & Murray46), its
inclusion in a package of measures meant that confounding effects could not be discounted. The same
criticism applied to the fourth intervention, remuneration policy. While a US study claimed that
incremental increases in driver pay led to slightly larger decreases in crash risk, there was no reported
control for other changes that were introduced at the same time.
The recent review by Newnam & Watson22 refers to intervention strategies ‘commonly adopted by
proactive organisations’, but gives no information about any evaluation of these strategies. It may be
noted that neither of these reviews mentions the topic of safety climate, or safety culture, which may
seem surprising, given the frequency with which these terms have appeared in the literature over the
last decade. One possible explanation is that safety climate/culture has tended to be seen as a
desirable attribute of an organisation, rather than as an intervention in its own right. Only one study
could be found in the WRRS literature that has investigated the relationship between safety climate
and safety; this was an early exploratory study by Wills et al.47 The results showed that there were no
significant differences in fleet safety climate scores between drivers who had been involved in
accidents and those who had not. Later work by the same authors48 has focused on driver behaviour
or, rather, self-reported driver behaviour.
It would appear that the evaluation literature – such as it is – offers little in the way of support for
interventions at the organisational level to improve WRRS. This conclusion may seem at first sight to be
at variance with a number of case studies that have claimed marked improvements in safety outcomes as
a result of instituting changes in organisational policies and procedures. This issue will be discussed in the
next section. The results from the review of the evaluation studies are summarised in Table 1.
Table 1
Study Type Participants Measures
Summary of results
from intervention A. Driver training
studies (continued
on pages 21–23) Lynn & Lockwood11 Cross-sectional survey 2,417 ‘regular’ business Accident liability over 3
drivers years
Herz 1995, cited in Case study Half of fleet drivers Crash rates
Haworth et al.19
Gregersen et al.26 Quasi-experiment 988 drivers in test group, Accidents per 10,000 km
988 in control over 2 years
Salminen28 Before/after; no control 179 drivers (including External safety audit,

‘volunteers’) accident numbers
Darby et al.29 Retrospective analysis 378 drivers Claims per year
B Group discussions
Ludwig & Geller,33 cited in Before/after; test/control Varying numbers Seat belt usage
Haworth et al.19
Salminen28 Before/after; no control 172 drivers Accident numbers over 3

years
C Incentives
Schneider36 Case study Not stated Accidents per 100,000 km

over 30 years
Lynn & Lockwood11 Cross-sectional survey 2,417 ‘regular’ business Accident liability over 3
drivers years
Mejza et al. 2003, cited Survey Not stated Success?

in Banks et al.21
Newnam et al.38 Before/after; no control 24 fleet managers Attitude change
D Publicity
Ludwig & Geller,33 cited in Before/after; with controls Varying numbers Behaviour modification
Haworth et al.19
Confounds* Results Comments
Multivariate analysis – Trained drivers had an accident liability Rigorous data analysis, but result non-
fully controlled 8% lower than non-trained drivers significant; possible self-selection bias
No information 35% reduction in crash rates Low quality: no information on

matching groups, no statistics, no data
on long-term effects
Matched control group; Significant 40% reduction in accident High quality: quasi-experimental design
exposure in measure rate and analysis. Unconventional training
content
No information Significant improvement in audit scores; Low quality: no control group, main
21% increase in accidents outcome measure not validated, no
measure of exposure, accident numbers
small
Multifactor study Significant 64% reduction in claims rate Rigorous data analysis; only high-risk
after training drivers studied, no control group,
training only one of several interventions
exposure in measure rate and analysis
Multifactor studies Increase in seat belt wearing rates Short term effect; no control for
confounding factors
No information 72% decrease in traffic accidents; Low quality: no control group, no

15% decrease in other accidents measure of exposure, accident numbers
small (18 vs 5)
No information 75–85% reduction in accident rates Low quality: no control, incentive

scheme only one of several measures
(multifactor)
exposure in measure rate and analysis
Multivariate analysis – Drivers in incentive schemes had Rigorous data analysis, but result only
fully controlled accident liability 21% lower than those marginally significant (p = 0.09)
not in the schemes
No information Successful trucking companies used Low quality: no comparison with

incentive schemes unsuccessful companies
No information Incentives (via insurance premiums) had Theoretically sound, but outcome
no effect on managers’ attitudes measure lacks validity
Matched control group; Non-significant 22% increase in High quality: quasi-experimental design
exposure in measure accident rate and analysis
No information Increases in targeted behaviours Behaviour measures lack validity, no

statistical testing
* Confounds: whether confounding variables were considered.

Table 1
Study Type Participants Measures
continued
E. In-vehicle data recorders
Berlin police department Before/after; no control 380 vehicles Accident numbers

1998, cited in ETSC41
Wouters & Bos43 Quasi-experiment 3,100 vehicle years in 6 Accident numbers

fleets
Toledo & Lotan42 Before/after 33 car drivers ‘Risk index’ from

recorders
Toledo et al.45 Before/after 191 truck drivers ‘Risk index’, crashes
F. Organisational
Wills et al.47 Survey 323 drivers Behaviour, offences and

crashes
White & Murray46 Case study Not stated Collision rates over 3
years
Confounds Results Comments
No information 20% reduction in accidents after fitting Low quality: no control, no exposure
recorders measure
Matched controls Significant 20% reduction in accidents High quality: quasi-experimental design
over 6 fleets and analysis
No information Significant correlations between index Sound demonstration project rather

and crash history; feedback improved than evaluation
behaviour in the short term
No information Significant 38% reduction in crash rates No proper control, no change in at-fault
in the 7 months after installation crash rates
Multiple regression No relation between safety climate and Possible self-selection bias
analysis offences or crashes
No information 13–30% reduction in crash numbers Low quality: no control, no exposure

measure, multifactor approach
5 Methodological issues
5.1 Introduction
Reference was made in an earlier section to a Cochrane review of driver training, and to the fact that
such reviews are regarded as the ‘gold standard’ for methodological rigour. Having reviewed the
literature on WRRS interventions, it has to be concluded that it is very probable that there is not a
single item in it that would be deemed acceptable for inclusion in a Cochrane review. This is not as
damning as it sounds, for it is also very probable that only a very small part of the entire road safety
literature would be eligible. The basic reason is that randomised allocation – the central tenet of the
experimental method – is extremely difficult to achieve in the real world, and particularly in an
occupational or commercial setting. There are also numerous other barriers to achieving quality in
evaluation studies in this area. As Downs et al.12 noted:
At a practical level, it is possible that organisations are reluctant to disclose the scale of their
accident problem, or may regard the information as commercially sensitive and inappropriate for
publication. It is also possible that accident data is not always collected in a sufficiently reliable
and consistent way that would permit statistical analysis.
They then go on to note the difficulties in establishing proper control groups, the statistical problem
of regression to the mean, the possibility of selection bias, and the problem of potentially
confounding effects.
Without in any way underestimating these problems, they are not insuperable. Robson49 describes a
number of established procedures that are suitable for use in what he terms ‘real world research’, and
the Swedish study by Gregersen et al.26 showed that experimentation (or, more strictly, quasi-
experimentation) is a feasible proposition. Despite this, the evaluation scene for WRRS is a depressing
one. The recent study by Banks et al.21 identified 20 peer-reviewed articles covering 19 WRRS
initiatives. Of these, only six were found to be effective in the post-intervention period. The phrase
‘the article did not indicate if the changes in outcome were significant’ was used by the authors eight
times. On six further occasions, it was found that the intervention under consideration had been
introduced as part of a package of measures, and so its unique contribution could not be assessed.
The authors found only four instances in the last 20 years of a well-controlled study in which an
intervention resulted in a significant reduction in accident rates. This hardly seems to justify the
assertion made by Rowland et al.50 that ‘research has shown that appropriately designed, industry-
based road safety interventions can reduce the number and severity of work-related road incidents’.
In their conclusions, Banks et al.21 raise the possibility that their selection criteria may have been too
rigorous, and that including non-peer reviewed publications might have broadened the scope. This
approach was adopted in the review by Newnam & Watson;22 while it drew attention to a number of
research publications, it was not able to identify any additional studies that had assessed the
effectiveness of interventions. This must raise the possibility that evaluation research has somehow
gone out of fashion in WRRS.
5.2 The multidimensional approach

One very obvious feature of changing practices in the last decade has been the shift from single to
multiple interventions. In the predominantly Australian literature during this period, this has often
been justified by drawing marked distinctions between procedures and practices. These distinctions
sometimes seem more apparent than real, and merit further attention.
The first distinction is the one that has been drawn between the ‘silver bullet’ approach and the
multidimensional approach. Thus, Rowland et al.50 state that:
Historically in terms of exploring and implementing fleet safety interventions, organisations have
often taken a ‘silver bullet’ approach aimed at developing a single countermeasure or intervention
strategy to encompass and address all work-related road safety issues.
They later maintain that ‘past research has revealed that a single reactive approach is ineffective in
improving long term work-related vehicle/road safety’ but cite no evidence to support this position.
Their preferred option is that organisations should adopt a ‘proactive multiple strategy approach’.
Such advice leads inevitably to the methodological problems encountered by Banks et al.21 in their
review, when they held that ‘as many of the studies reviewed investigated the effects of initiatives
when implemented in combination with other initiatives, it was not possible to distinguish the unique
impact of some initiatives’. Another way of putting this might be to say that proper evaluation
becomes extremely difficult when the silver bullet is replaced by the blunderbuss.
A second distinction has been drawn between what is termed the ‘asset management approach’ and
its alternative, the ‘human behaviour interface’. Asset management is regarded as a reactive approach,
concerned only with the frequency and severity of fleet accidents (Freeman et al.51). It is criticised for
being simplistic and for not providing information on underlying attitudinal and behavioural factors
that can be gained by using the alternative approach. If this is a plea for better data collection, it is
commendable; otherwise, the distinction is a somewhat artificial one.
The third distinction that has been made is between ‘data-driven’ and ‘theory-led’ approaches to
WRRS (Newnam & Watson22). The argument set out by these authors is that:
Although anecdotal and data-driven research has provided significant insight into the extent of
the work-related driving problem, it is limited in that it does not provide the theoretical
foundation for establishing the development and improvement of interventions designed to reduce
death and injury in the work vehicle.
Leaving aside the question of whether anecdotal data can really provide ‘significant’ insight into this
problem, it is not unreasonable to ask why effective interventions can only be achieved by applying
top–down theory. The study by Broughton et al.52 is relevant here. Unconstrained by theoretical
preconceptions, it employed rigorous data analysis in order to identify aspects of behaviour that were
significantly associated with accident risk, and thereby propose possible interventions to improve
WRRS.
5.3 Intervening variables

According to Freeman et al.,51 ‘… practitioners and researchers are beginning to direct an increasing
level of focus towards developing and utilising self-report measurement scales to determine whether
practical relationships exist between self-reported attitudes and behaviours as well as subsequent
crash involvement’. They add that ‘one of the most prominent reasons for this focus is that vehicle
accidents have been found to be attributable to employees’ attitudes and behaviours’. In support of
this contention they cite the study by Chapman et al.,14 which is somewhat surprising, given that the
study in question did not include any attitudinal measures. Nonetheless, an increasing amount of
research in the last decade, mainly from Australasia but also from the UK, has been concerned with
the investigation of attitudinal and behavioural factors associated with crash involvement. The
rationale for this activity is that, if it can be shown that self-report measures are predictive of crash
involvement, then they would provide a richer and more informative outcome measure than the mere
counting of crash frequencies – which is the asset management approach referred to above.
Implementations that have been employed include the Driver Behaviour Questionnaire (DBQ), the
Driver Attitude Questionnaire (DAQ), the Driver Stress Index (DSI), the Safety Climate Questionnaire
(SCQ), as well as scales measuring risk taking and driving skills. The most widely used of these tools
has been the DBQ, and claims have been made that its measures are significantly associated with, and
predictive of crash involvement in, a fleet setting, eg Sullman et al.,53 Xie & Parker54 and Rowland et
al.55 ‘Prediction’ is a complex issue, and some authors (eg Freeman et al.51) have expressed
reservations about the effectiveness of this approach. A number of issues of both theoretical and
empirical concern have been raised:
• the DBQ was developed and, more importantly, standardised on the general driving population
• when used in a fleet driving context, its factor structure has been inconsistent
• studies have tended to show that the DBQ can only explain relatively small proportions of the
variance in crash involvement
• association does not imply causation
• the fact that a measure contributes to a predictive equation does not mean that it can act as a
proxy measure for the dependent variable in question.
There are two empirical studies that are of particular relevance here. The first is by Wåhlberg et al.,3
which looked at the predictive value of the DBQ, and concluded that ‘it may not be as successful in
predicting accidents as is often claimed’. The second is the study by Broughton et al.,52 in which
multivariate analysis found that reported violations as measured by the DBQ did not contribute to
their models of accident liability (a finding that has been little reported in the literature). A realistic
assessment should be that the prospect for developing proxy measures for fleet accidents does not at
present look particularly promising.
5.4 Case studies

If evaluation studies appear to be out of fashion in the WRRS field, then one type of investigation
that is in vogue is the case study. A number of these have appeared in the literature in recent times (eg
Lang et al.,8 White & Murray46 and Murray et al.56). They typically describe success stories in
organisations, and have given considerable encouragement to those who wish to adopt the ‘best
practice’ approach. Robson49 writes approvingly of case studies as a way of understanding social
change, and is used as a model by Murray et al.56 There can be no doubt that some of these case
study outcomes are very impressive. The Murray et al. study describes in detail the changes put into
practice by a major company with the aim of improving its WRRS, and cites as evidence a halving of
its collision rate and the fact that it ‘gained wider benefits by adopting a holistic approach’.
However, these success stories need to be put into a wider context. Where fleet safety is concerned,
the majority of positive outcomes relate to large organisations that are well motivated and have
shown a commitment to achieve and maintain improvements in WRRS. The picture painted by
Murray et al.57 in their study of small enterprises is a very different one, when they stated that:
Crash outcomes were purposefully not included as one of the initial project targets… [because]
participants would be secretive about such data… It was better to concentrate on more proactive
outcomes, particularly the implementation of processes and systems.
Practitioners readily point to case studies as clear examples of how the organisational approach to
fleet safety can achieve positive results. However, there are two issues that should be borne in mind
regarding case studies. First, as noted above, only positive outcomes tend to be reported. There is
anecdotal evidence that some large organisations have implemented programmes without achieving
any benefits. If the ratio of successful to unsuccessful programmes is unknown, then the value of such
interventions must remain unclear. Second, there is the problem of generalisation. Case studies are a
very good method of showing what can be achieved by bringing about change, but they provide only
limited information on how these results can be transferred to other situations. One recurring theme
in the literature has been the disparate nature of commercial fleets, and this very fact tends to weaken
the predictive validity of case studies.
5.5 Barriers to evaluation

This section considers methodological problems that have impinged on WRRS in the past, continue
to do so in the present, and act as barriers to evaluation. The first, and most obvious, is that it takes
place in the real world, and therefore experimentation to establish the effectiveness of remedial
measures is extremely difficult. Twenty years ago, this was seen as a challenge; good evaluation work
was carried out, and possibilities for further research were identified. These possibilities have never
been followed up. Largely as a result of pressure from policy makers and legislators, there has been a
major change in emphasis where a concern with policies and procedures has become the main
priority for fleet operators. While Murray et al.20 stress the need for ‘proactive key performance
indicators’, they also accept that much of the activity in the UK in the last decade has been concerned
with identifying and disseminating best practice, rather than with evaluation. They also go on to
argue that:
Even where safety gains in a pure academic or statistical sense cannot be proved, a safety
programme has very high face validity, helps provide protection from regulations … and offers a
range of PR and business development opportunities.
Those convinced by this argument could well be led to question the need to invest time and resources
in evaluation research that could have uncertain outcomes.
The second barrier is the near universal use of multifactor interventions. While case study evidence is
highly supportive of such an approach, what is lacking is an understanding of the relative
effectiveness of the component parts of any intervention package. Concern with procedure in the UK
has often led to multifactor interventions being adopted on a largely pragmatic basis. This contrasts
with the situation in Australia, where the multifactor approach is employed with enthusiasm and
buttressed by a range of theoretical and conceptual frameworks. In the UK, it would seem that the
concern is more with action than research; in Australia, it appears to be more with research than
evaluation.
Another barrier to evaluation has been the lure of intervening variables. To most road safety
researchers, accident statistics are the ultimate criterion by which any safety countermeasure should
be assessed. In the WRRS field, however, there has been increasing pressure to move to other more
‘explanatory’ measures, such as behaviour and attitudes, partly because of the ease with which such
data can be collected. There are a number of inherent problems with this approach, and it has
achieved only limited success to date.
The final barrier to evaluation is that it is no longer seen as being important – at least not in the
rigorous scientific sense. In the multifactor approach, changes to crash rates are now seen as only one
of many possible key performance indicators. As an example, in the Wolseley case study by Murray et
al.56 there were 11 other measures cited as outcomes in addition to a reduction in collision rate, the
majority of them non-quantifiable.
One effect of these barriers to evaluation is the fact that not a single properly controlled evaluation
study of WRRS interventions has been carried out in the last decade.
5.6 The content of interventions

One methodological issue that has not been addressed to any real extent in the literature is that of the
formulation and content of intervention measures. ‘Methodology’ is all too often seen as being only
concerned with experimental design and statistical analysis. In its proper sense, methodology is the
study of the principle guiding the search for knowledge, and as such is also concerned with problem
identification and the selection of the most appropriate means by which to address problems. Most of
the literature that has been discussed so far has been concerned only with the application of
intervention measures; with a few exceptions, there has been little in the way of assessing how
appropriate these measures are to the problem in hand, or, more importantly, of attempts to gain a
better understanding of the problem itself.
Earlier in the review, it was noted that the 1990s ended with a good understanding of the fleet driver
problem, but with no real evidence that could guide the design of interventions to improve WRRS.
Little has changed in the intervening period. Although Newnam & Watson22 argue that research in
this area has been data-driven, this is difficult to sustain. A careful examination of the literature
should lead to the clear conclusion that WRRS is in fact data-poor, particularly in respect of an
understanding of the basic problem and the factors that are associated with it.
National authorities faced with a rising accident problem spend time and resources to identify the
areas of greatest risk so that intervention measures could be deployed most effectively. By contrast,
the WRRS response has tended to be either to employ the ‘silver bullet’ of driver training, or else to
engage in multifactor organisational approaches. Even when problems have been well identified, the
findings have sometimes been discounted. For example, Wishart et al.58 identified seven main crash
types in their research on fleets, but held that such categorisation is merely part of their disparaged
‘asset management’ approach, which, they maintain, can only be concerned with short-term financial
gain rather then providing the information for what is really needed – ‘large scale behavioural
intervention and workplace culture change’. This is a clear example of theory taking precedence over
data.
The traditional road safety approach would be to employ the procedures of accident analysis and risk
factor identification to look in detail at the problem areas, and then direct interventions to the
optimal effect. The key issue is the need for a good understanding of the problem through the
collection of good quality reliable data. It may be that this does happen in some enlightened
organisations, but it does not appear in the literature.
However, there is some evidence that can contribute to this debate. Safety measures are most effective
when directed at high risk targets, so the identification of risk factors should be central to the
development of effective countermeasures. At a subjective level, Salminen & Lähdeniemi59 sent a
questionnaire to some 5,000 occupational drivers that included nine ‘risk factors’. Responses showed
that time pressure, tiredness and use of mobile phones were selected most as being factors that were
felt to increase risks while driving.
Robb et al.60 carried out a Cochrane-style systematic review of risk factors for work-related crashes
and injuries. While noting that methodologically sound studies were relatively rare in the literature,
the authors pointed out that the most consistent evidence related to fatigue and sleep-related factors,
and suggested that these could be major causes of work-related traffic injuries.
In France, Fort et al.61 used a case-control design to investigate the risk factors associated with
occupational road accidents. They identified a number of factors relating to scheduling and fatigue
management as contributing to increased risk of accidents.
Finally, Broughton et al.52 included a number of potential risk factors in their multivariate analysis of
work-related injury accidents. They also identified fatigue-related factors, time pressures and in-car
distractions such as mobile phones as significantly increasing the risk of accidents.
Thus, four studies using different procedures arrived at very similar conclusions. If one were to look
for a unifying concept, the most obvious candidate would be that of driver stress. A considerable
body of theoretical and practical work has been carried out on this topic in the last two decades, and
increasingly attention is being paid to the issue of driver stress in the occupational setting (see for
example Dorn et al.62 and Öz et al.63). It seems clear that the data-driven approach can not only
provide insight but also draw upon wider research evidence for guidance on how and where to target
interventions most effectively.
6 Summary
This study set out to provide evidence-based advice to practitioners about the effectiveness of
interventions to improve WRRS. This has proved to be a surprisingly difficult task, largely due to the
nature of the literature. Since 1999, there have been six reviews of the literature. All six experienced
difficulty in finding well-controlled evaluation studies. If one adopts the not unreasonable criterion
that an evaluation study should assess whether an intervention has brought about a statistically
reliable change in crash rates, then the results are meagre in the extreme. Only four interventions
meet this criterion, three were in the same investigation, and all were conducted more than a decade
ago.
In 1996, Gregersen et al.26 described the results of a major investigation in which four interventions
were compared, together with a control. Three of the interventions were shown to have a significant
effect in reducing accident rates. In 2000, Wouters & Bos43 were able to show that the installation of
‘black box’ recorders could reduce accident involvement among fleet drivers. These four interventions
are the only ones in the literature that show scientific credibility. In the last decade, there have been
no serious evaluation studies undertaken, and the claims of improvement in WRRS have been largely
anecdotal in nature.
It is important to bear in mind that the absence of strong evidence of effectiveness does not mean that
effectiveness has not been achieved. There have been a number of case studies in recent times that
have claimed impressive gains in WRRS through the implementation of large scale programmes. Two
notes of caution must be sounded, however. The first is that it is in the nature of things that only
success stories get reported. There is some anecdotal evidence that programmes have been carried out
in some large organisations without making any impact on existing problems. The second is that the
current practice is to use a broad package of measures, which means that the effectiveness of each
individual component is therefore impossible to assess. Without a return to ‘classical’ evaluation,
work to improve WRRS can only be on an ad hoc and less than efficient basis. There are of course
commercial and practical issues that remain to be overcome when trying to persuade industry to
engage in evaluation studies. Companies may not wish to invest time and money in evaluation studies
if they perceive that holistic approaches (even if only supported by case study data) are sufficient for
their needs. In addition, companies may be reluctant to have their accident data made public.
Nonetheless, the fact remains that there is a pressing need to use controlled evaluation studies to
assess which WRRS interventions work, by how much, and through which causal mechanisms.
References
1 World Health Organization. Global status report on road safety. Time for action. WHO, 2009.
http://whqlibdoc.who.int/publications/2009/9789241563840_eng.pdf (accessed 11 September
2009).
2 Work-related Road Safety Task Group. Reducing at-work road traffic incidents. HSE Books,
2001.
3 Wåhlberg A, Dorn L and Kline T. The Manchester Driver Behaviour Questionnaire as a
predictor of road traffic accidents. Theoretical Issues in Ergonomics Science 2011; 12 (1): 1–21.
4 Boufous S, Ivers R, Senserrick T, Stevenson M, Norton R and Williamson A. Accuracy of self-
report of on-road crashes and traffic offences in a cohort of young drivers: the DRIVE study.
Injury Prevention 2010; 16: 275–277.
5 Arthur W, Bell S, Edwards B, Day E, Tubre T and Tubre A. Convergence of self-report and
archival crash involvement data: a two year longitudinal follow-up. Human Factors 2005; 47
(2): 303–313.
6 Maycock G, Lockwood C and Lester J. The accident liability of car drivers. TRRL Report
RR315. Transport and Road Research Laboratory, 1991.
7 Coolican H. Research methods and statistics in psychology. Hodder & Stoughton, 2004.
8 Lang B, Delmonte E and Vandrevala T. Development and evaluation of the Work Related Road
Safety CD-ROM. Published Project Report 346. Transport Research Laboratory, 2009.
9 Murray W. Worldwide Occupational Road Safety (WORS) Review Project. National Institute
for Occupational Safety and Health, 2007.
10 European Transport Safety Council. Reducing road safety risk driving for work and to work in
the EU. ETSC Position Paper, February 2010. ETSC, 2010.
11 Lynn P and Lockwood C. The accident liability of company car drivers. TRL Report 317.
Transport Research Laboratory, 1998.
12 Downs C G, Keigan M, Maycock G and Grayson G B. The safety of fleet car drivers: a review.
TRL Report 390. Transport Research Laboratory, 1999.
13 Dimmer A R and Parker D. The accidents, attitudes and behaviour of company car drivers. In
Grayson G B (ed.) Behavioural Research in Road Safety IX. Transport Research Laboratory, 1999.
14 Chapman P, Roberts K and Underwood G. A study of the accidents and behaviours of company
car drivers. Behavioural Research in Road Safety X. Department of Transport, 2000.
15 Grayson G B. Company cars and road safety. In Grayson G B (ed.) Behavioural Research in
Road Safety IX. Transport Research Laboratory, 1999.
16 Lang B and Rehm L. Literature review on van use in the UK. Published Project Report 113.
Transport Research Laboratory, 2006.
17 Health and Safety Executive. Driving at work: managing work-related road safety, INDG382.
HSE/Department for Transport, 2003. www.hse.gov.uk/pubns/indg382.pdf (accessed 28 June
2011).
18 Bibbings R. Occupational road risk: towards a management approach. Journal of the Institution
of Occupational Safety and Health 1997; 1 (1): 61–75.
19 Haworth N, Tingwall C and Kowaldo N. Review of best practice fleet safety initiatives in the
corporate and/or business environment. MUARC Report 166. Monash University Accident
Research Centre, 2000.
20 Murray W, Newman S, Watson B, Davey J and Schonfeld C. Evaluating and improving fleet
safety in Australia. ATSB Report, November 2002. Australian Transport Safety Bureau, 2003.
21 Banks T, Davey J, Biggs H and King M. A review of the effectiveness of occupational road
safety initiatives. In Dorn L (ed.) Driver behaviour and training: Volume IV. Ashgate, 2010.
22 Newnam S and Watson B. Work-related driving safety in light vehicle fleets: a review of past
research and the development of an intervention framework. Safety Science 2011; 49 (3):
369–381.
23 Helman S, Grayson G B and Parkes A. How can we produce safer new drivers? A review of the
effects of experience, training and limiting exposure on the collision risk of new drivers. TRL
Report INS005. Transport Research Laboratory, 2010.
24 Lund A K and Williams A F. A review of the literature evaluating the Defensive Driving Course.
Accident Analysis & Prevention 1985; 17: 449–460.
25 Ker K, Roberts I, Collier T, Beyer F, Bunn F and Frost C. Post-licence driver education for the
prevention of road traffic crashes. Cochrane Database of Systematic Reviews Issue 3, 2003. Art
no: CD003734.
26 Gregersen N P, Brehmer B and Morén B. Road safety improvement in large companies. An
experimental comparison of different measures. Accident Analysis & Prevention 1996; 28: 297–306.
27 Gregersen N P. Company car drivers – VTI work. In Grayson G B (ed.) Behavioural Research in
Road Safety IX. Transport Research Laboratory, 1999.
28 Salminen S. Two interventions for the prevention of work-related road accidents. Safety Science
2008; 46: 545–550.
29 Darby P, Quddus M, Murray W, Raeside R and Ison S. Evaluation of fleet road safety
interventions. Paper presented at the 90th Annual Meeting of the Transportation Research
Board, Washington DC, 2011.
30 European Transport Safety Council. Fit for road safety: from risk assessment to training.
PRAISE Report No 2. ETSC, 2010.
31 Lewin K. Frontiers in group dynamics. Human Relations 1947; 1: 5–38.
32 Misumi J. The effects of organisational climate variables, particularly leadership variables and
group decisions, on accident prevention. Paper given to the 19th International Congress of
Applied Psychology, Munich, 1978.
33 Ludvig T and Geller E S. Intervening to improve the safety of occupational driving: a behaviour
change model and review of empirical evidence. Journal of Organizational Behavior
Management 2000; 19 (4): 1–124.
34 Wilde G J S. Target risk. PDE Publications, 1994.
35 Wilde G J S and Murdoch P. Incentive systems for accident-free and violation-free driving in the
general population. Ergonomics 1982; 25: 879–890.
36 Schneider W. Influencing individual and group safety values and norms. In Koornstra M and
Christensen J (eds) Enforcement and rewarding: strategies and effects. SWOV, 1990.
37 Koornstra M J and Christensen J. Enforcement and rewarding: strategies and effects. SWOV, 1990.
38 Newnam S, Tay R and Mason C. Using psychological frameworks to inform the evaluation of
fleet safety initiatives. Safety Science 2006; 44: 809–820.
39 Ludvig T and Geller E S. Behavior change among agents of a community safety program: pizza
deliverers advocate community safety belt use. Journal of Organizational Behavior Management
1999; 19 (2): 3–24.
40 Larson L, Schnelle J, Kirchner R, Carr A, Domash M and Risley T. Reduction of police vehicle
accidents through mechanically aided supervision. Journal of Applied Behavior Analysis 1980;
13: 571–581.
41 European Transport Safety Council. How can in-vehicle safety equipment improve road safety
at work? PRAISE Report No 1. ETSC, 2009.
42 Toledo T and Lotan T. An in-vehicle data recorder for evaluation of driving behavior and safety.
Transportation Research Record 2006; 1953: 112–119.
43 Wouters P and Bos J M. Traffic accident reduction by monitoring driver behaviour with in-car
data recorders. Accident Analysis & Prevention 2000; 32: 643–650.
44 SAMOVAR. DRIVE Project V2007: final project report. Department of Electronic Engineering,
Queen Mary College, University of London, 1995.
45 Toledo T, Musicant O and Lotan T. In-vehicle data recorders for monitoring and feedback on
drivers’ behaviour. Transportation Research Part C 2008; 11: 320–331.
46 White J and Murray W. Occupational road safety case study: Roche Australia cuts risks,
collisions and costs. Journal of Australasian College of Road Safety 2007; 18 (3): 28–29.
47 Wills A R, Watson B and Biggs H C. The relative influence of fleet safety climate on work-
related driver safety. Proceedings of the Australasian Road Safety Research Conference, Perth,
Australia, 2004.
48 Wills A R, Watson B and Biggs H C. Comparing safety climate factors as predictors of work-
related driving behaviour. Journal of Safety Research 2006; 37: 375–383.
49 Robson C. Real world research. Blackwell, 1993.
50 Rowland B, Wishart D and Davey J. Occupational fleet safety research: a case study approach.
Proceedings of the 13th Annual Occupational Health and Safety Conference, Cairns, 2005.
51 Freeman J, Wishart D, Davey J and Rowland B. Developing risk-assessment tools for fleet
settings: where to from here? In Dorn L (ed.) Driver behaviour and training: Volume IV.
Ashgate, 2010.
52 Broughton J, Baughan C, Pearce L, Smith L and Buckle G. Work-related road accidents. TRL
Report 582. Transport Research Laboratory, 2003.
53 Sullman, M Meadows M and Pajo P. Aberrant driving behaviours amongst New Zealand truck
drivers. Transportation Research Part F 2007; 10: 217–232.
54 Xie C and Parker D. A social psychological approach to driving violations in two Chinese cities.
Transportation Research Part F 2007; 10: 293–308.
55 Rowland B, Davey J, Freeman J and Wishart D. Work-related road safety risk assessment:
utilisation of self-report surveys to predict organisational risk. Proceedings of Australasian Road
Safety Research Conference, Adelaide, 2008.
56 Murray W, Ison S, Gallemore P and Nijjar H S. Effective occupational road safety programs: a
case study of Wolseley. Transportation Research Record 2009; 2096: 55–64.
57 Murray W, Bamford C and Whiteing T. Implementing a fleet safety programme in UK-based
small and medium sized enterprises. Proceedings of RoSPA 67th Road Safety Congress,
Stratford upon Avon, 2002.
58 Wishart D, Davey J and Freeman J. A review of developing and implementing Australian fleet
safety interventions: a case study approach update. In Dorn L (ed.) Driver hehaviour and
training: Volune III. Ashgate, 2008.
59 Salminen S and Lähdeniemi E. Risk factors in work-related traffic. Transportation Research
Part F 2007; 10: 77–86.
60 Robb G, Sultana S, Ameratunga S and Jackson R. A systematic review of epidemiological
studies investigating risk factors for work-related road traffic crashes and injuries. Injury
Prevention 2008; 14: 51–58.
61 Fort E, Pourcel L, Davezies P, Renaux C, Chiron M and Charbotel B. Road accidents, an
occupational risk. Safety Science 2010; 48 (10): 1412–1420.
62 Dorn L, Stephen L, Wåhlberg A and Gandolfi J. Developing and validating a self-report
measure of bus driver behaviour. Ergonomics 2010; 53 (12): 1420–1433.
63 Öz B, Özkan T and Lajunen T. Professional and non-professional drivers’ stress reactions and
risky driving. Transportation Research Part F 2010; 13: 32–40.
This document is printed on chlorine-free paper produced from managed, sustained forests.
IOSH IOSH is the Chartered body for health and safety
The Grange professionals. With more than 39,000 members
Highfield Drive in 85 countries, we’re the world’s largest
Wigston professional health and safety organisation.
Leicestershire
LE18 1NN We set standards, and support, develop and
UK connect our members with resources, guidance,
events and training. We’re the voice of the
t +44 (0)116 257 3100 profession, and campaign on issues that affect
f +44 (0)116 257 3101 millions of working people.
www.iosh.co.uk
IOSH was founded in 1945 and is a registered
charity with international NGO status.
AC/11191/MO/110711/P
£15
Institution of Occupational Safety and Health

Founded 1945
Incorporated by Royal Charter 2003
Registered charity 1096790

Work Related Road Safety Literature Review

Uploaded by

Copyright:

Available Formats

Work Related Road Safety Literature Review

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Work Related Road Safety Literature Review

Uploaded by

Copyright:

Available Formats

Work-related

Report submitted to the IOSH Research Committee

Graham B Grayson BSc PhD and Shaun Helman BSc PhD

Report submitted to the IOSH Research Committee

Graham B Grayson BSc PhD and Shaun Helman BSc PhD

4 Interventions to improve work-related road safety 14

In-vehicle data recorders (IVDRs)

Experimentation to establish the effectiveness of remedial measures is extremely difficult. Twenty

1.1 The problem

1.2 Why this review is needed

2.1 Inclusion criteria

2.2 Search methods and terms

2.3 Data collection

3.2 The past (1970s to mid-1990s)

3.3 The present (mid-1990s to present day)

• a requirement for safe plants and their maintenance

4 Interventions to improve work-related road safety

4.2 Driver training

As will be seen, practice has proved to be very different.

4.3 Group discussions

• a 60-minute ‘warm-up’ period of free discussion

4.6 In-vehicle data recorders (IVDRs)

4.7 Organisational approaches

Salminen28 Before/after; no control 179 drivers (including External safety audit,

Darby et al.29 Retrospective analysis 378 drivers Claims per year

Salminen28 Before/after; no control 172 drivers Accident numbers over 3

Schneider36 Case study Not stated Accidents per 100,000 km

Mejza et al. 2003, cited Survey Not stated Success?

Newnam et al.38 Before/after; no control 24 fleet managers Attitude change

Confounds* Results Comments

No information 35% reduction in crash rates Low quality: no information on

No information 72% decrease in traffic accidents; Low quality: no control group, no

No information 75–85% reduction in accident rates Low quality: no control, incentive

No information Successful trucking companies used Low quality: no comparison with

No information Increases in targeted behaviours Behaviour measures lack validity, no

* Confounds: whether confounding variables were considered.

Berlin police department Before/after; no control 380 vehicles Accident numbers

Wouters & Bos43 Quasi-experiment 3,100 vehicle years in 6 Accident numbers

Toledo & Lotan42 Before/after 33 car drivers ‘Risk index’ from

Toledo et al.45 Before/after 191 truck drivers ‘Risk index’, crashes

Wills et al.47 Survey 323 drivers Behaviour, offences and

Confounds Results Comments

No information Significant correlations between index Sound demonstration project rather

No information 13–30% reduction in crash numbers Low quality: no control, no exposure

5.2 The multidimensional approach

5.3 Intervening variables

5.4 Case studies

5.5 Barriers to evaluation

5.6 The content of interventions

Institution of Occupational Safety and Health

You might also like