aeysrtge o rural roads safety strategies for

More than 75 000 people are killed each year on rural roads in OECD countries; this represents about 60 per cent of fatal road crashes. This loss ROAD TRANSPORT AND INTERMODAL RESEARCH of lives has an economic cost of around US$135 billion per year. The relative share of rural road fatalities in total road fatalities has risen from less than 55 per cent in 1980 to more than 60 per cent in 1996. While there has been a reduction in the total number of road fatalities in OECD countries during the past 20 years, it is clear that safety improvements on motorways and urban roads have been more successful than those on rural roads.

Following an in-depth review of the characteristics of road crashes in rural areas, the book proposes a series of safety measures, focusing on infrastructure management, enforcement, innovative tools, such as intelligent transport systems, and trauma management. The book strives to raise the awareness of road users, decision makers and politicians of the importance of road safety in rural areas. It is a very useful handbook for local, regional or national authorities seeking to improve – sometimes at very low cost – safety on rural roads.

safety strategies

OECD for rural roads

(77 1999 01 1 P) FF 240 9:HSTCQE=V\UZYY: ISBN 92-64-17054-5 -99 ROAD TRANSPORT AND INTERMODAL RESEARCH

Safety Strategies for Rural Roads

ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVELOPMENT

Pursuant to Article 1 of the Convention signed in Paris on 14th December 1960, and which came into force on 30th September 1961, the Organisation for Economic Co-operation and Development (OECD) shall promote policies designed: ± to achieve the highest sustainable economic growth and employment and a rising standard of living in Member countries, while maintaining ®nancial stability, and thus to contribute to the development of the world economy; ± to contribute to sound economic expansion in Member as well as non-member countries in the process of economic development; and ± to contribute to the expansion of world trade on a multilateral, non-discriminatory basis in accordance with international obligations. The original Member countries of the OECD are Austria, Belgium, Canada, Denmark, France, Germany, Greece, Iceland, Ireland, Italy, Luxembourg, the Netherlands, Norway, Portugal, Spain, Sweden, Switzerland, Turkey, the United Kingdom and the United States. The following countries became Members subsequently through accession at the dates indicated hereafter: Japan (28th April 1964), Finland (28th January 1969), Australia (7th June 1971), New Zealand (29th May 1973), Mexico (18th May 1994), the Czech Republic (21st December 1995), Hungary (7th May 1996), Poland (22nd November 1996) and Korea (12th Decem- ber 1996). The Commission of the European Communities takes part in the work of the OECD (Article 13 of the OECD Convention).

PubliÂe en franËcais sous le titre : STRATEGIESÂ DE SECURITÂ EÂ ROUTIEREÁ EN RASE CAMPAGNE

 OECD 1999 Permission to reproduce a portion of this work for non-commercial purposes or classroom use should be obtained through the Centre franËcais d'exploitation du droit de copie (CFC), 20, rue des Grands-Augustins, 75006 Paris, France, Tel. (33-1) 44 07 47 70, Fax (33-1) 46 34 67 19, for every country except the United States. In the United States permission should be obtained through the Copyright Clearance Center, Customer Service, (508)750-8400, 222 Rosewood Drive, Danvers, MA 01923 USA, or CCC Online: http://www.copyright.com/. All other applications for permission to reproduce or translate all or part of this book should be made to OECD Publications, 2, rue AndrÂe-Pascal, 75775 Paris Cedex 16, France. FOREWORD

The Road Transport and Intermodal Linkages Research Programme (RTR) is a co-operative approach among Member countries to address technical, economic and policy issues relevant to safe and efficient road transport. The Programme, through its broader linkages to other modes, reflects a multimodal approach to common transport problems and represents a combined attempt to reduce the negative impact of transport on the environment. The Programme has two main fields of activity:

• International research and policy assessments of road and road transport issues to provide analytical support for decisions by Member governments and international governmental organisations;

• Technology transfer and information exchange through two databases -- the International Road Research Documentation (IRRD) scheme and the International Road Traffic and Accident Database (IRTAD).

Its mission is to:

• enhance innovative research through international co-operation and networking; • undertake joint policy analyses and prepare technology reviews of critical road transport issues; • promote the exchange of economic, scientific and technical information in the transport sector and contribute to road technology transfer in OECD Member and non-member countries; • promote the development of sound policies to achieve a safe and efficient transport sector that is responsive to the environment.

The activities concern:

• sustainable multimodal transport strategies; • economic performance, transport infrastructure and management; • transport safety and environment.

3 ABSTRACT

IRRD NO 491006

Each year, more than 75 000 people are killed on rural roads in OECD Member countries. These deaths are accompanied by economic costs on the order of USD 135 billion per year. The relative importance of rural road fatalities in relation to total road fatalities has climbed from less than 55% in 1980 to more than 60% in 1996. Because OECD countries have experienced a reduction in the total number of road crash fatalities, it is clear that motorway and urban road safety improvements have been more successful than those on rural roads. The OECD therefore created an Expert Group composed of representatives from 13 countries to examine the problems and propose strategies for improving the situation. The report shows that as much as 80% of all accidents on rural roads fall into three categories: single vehicle accidents, head-on collisions and collisions at intersections. A main conclusion from this information is that the rural road system itself has inherent characteristics that significantly contribute to the high number of accidents and the high risks. It is therefore recommended that every OECD Member country should develop a rural road safety improvement strategy. It is also recommended that each country should develop short-, short-/medium- and long-term programmes. Such plans must pay special attention to raising awareness about rural road safety both within the general public and within the organisations of all key actors. Various safety measures that can improve rural road safety are suggested throughout the report. Although a network-wide approach is required and recommended in the report, it is understood that individual low-cost measures can contribute to the safety of the rural road network. The report recommends that safety should receive explicit attention at every level of the process, from the decision to build or rebuild, through planning, design and construction and during operation and maintenance. As there is currently insufficient information available on rural road safety problems to adequately support appropriate policy and investment decisions, the report recommends more systematic evaluation of the effectiveness of countermeasures. In addition to this, further research into rural road safety strategies and various individual safety measures is required to rapidly advance safety improvements on rural roads.

Field Classification: accident studies, accidents and the road.

Field Codes: 80, 82

Key Words: OECD, rural area, accident rate, fatality, accident prevention, policy, evaluation (assessment), danger, highway, improvement, efficiency.

4 TABLE OF CONTENTS

EXECUTIVE SUMMARY...... 9

Chapter I INTRODUCTION ...... 13 I.1 Defining rural roads...... 13 I.2 Rural road safety...... 14 I.3 Safety on rural roads is declining ...... 15 I.4 Safety policy for rural roads ...... 16 I.5 Objectives of the study ...... 16 I.6 Structure of the report...... 17 Chapter II GENERAL CHARACTERISTICS OF RURAL ROAD SAFETY ...... 19 II.1 Introduction...... 19 II.2 The rural road network and its use ...... 20 II.3 Speed limits and driving speeds ...... 21 II.4 Accidents and fatalities...... 22 II.5 The socio-economic cost ...... 27 II.6 Safety policy ...... 28 II.7 Conclusion ...... 29 BIBLIOGRAPHY ...... 31

Chapter III FACTORS CONTRIBUTING TO RURAL ROAD ACCIDENTS...... 33 III.1 Introduction...... 33 III.2 Review of the systemic nature of an accident ...... 33 III.3 The driver...... 34 III.4 The road environment...... 37 III.5 Vehicle/user characteristics ...... 41 III.6 Conclusion ...... 42 BIBLIOGRAPHY ...... 44

Chapter IV NETWORK PLANNING AND ROAD INFRASTRUCTURE...... 47 IV.1 Introduction...... 47 IV.2 Road traffic engineering and its potential for improving road safety ...... 47 IV.3 Network planning ...... 48 IV.4 Safety-related road design characteristics ...... 50 IV.5 Roadside safety...... 58 IV.6 Signs, markings and lighting ...... 62

5 IV.7 Maintenance and work zones...... 65 IV.8 Vulnerable road users: pedestrians and bicyclists...... 67 IV.9 Identifying (potential) safety problems and solutions...... 69 IV.10 Conclusion ...... 72 BIBLIOGRAPHY ...... 74

Chapter V ENFORCEMENT...... 79 V.1 Introduction...... 79 V.2 Effectiveness and limitations of enforcement ...... 79 V.3 The enforcement mechanism...... 82 V.4 Enforcement practices ...... 86 V.5 Management, priority setting and funding ...... 92 V.6 Conclusion ...... 93 BIBLIOGRAPHY ...... 94

Chapter VI POTENTIAL SOLUTIONS FROM INTELLIGENT TRANSPORT SYSTEMS...... 97 VI.1 Introduction...... 97 VI.2 General safety considerations of ITS...... 98 VI.3 Speed-control devices...... 98 VI.4 Driver/vehicle information systems ...... 101 VI.5 Infrastructure-based applications for co-operative systems ...... 104 VI.6 Human factor considerations ...... 105 VI.7 ITS in perspective...... 105 VI.8 Future needs...... 106 VI.9 Conclusion ...... 107 BIBLIOGRAPHY ...... 108

Chapter VII TRAUMA MANAGEMENT IN RURAL AREAS...... 109 VII.1 Special risks of rural road crashes ...... 109 VII.2 Timeliness of treatment ...... 109 VII.3 Road trauma treatment in rural environments...... 111 VII.4 Opportunities for improved trauma management in rural areas...... 112 VII.5 Conclusion ...... 114 BIBLIOGRAPHY ...... 115

Chapter VIII STRATEGIC FRAMEWORK...... 117 VIII.1 Introduction...... 117 VIII.2 The rural road network: a functional approach ...... 117 VIII.3 Rationalising policy making...... 119 VIII.4 Policy: organisation -- financing -- information/data...... 126 VIII.5 Summary...... 128 BIBLIOGRAPHY ...... 129

6 Chapter IX CONCLUSION AND RECOMMENDATIONS ...... 131 IX.1 Rural road safety -- a sleeping giant...... 131 IX.2 Strategy to improve rural road safety ...... 133 IX.3 Research needs...... 137 IX.4 Next steps -- dissemination of the results...... 138 Annex A LIST OF MEMBERS...... 139

7

EXECUTIVE SUMMARY

The rural road safety problem is serious

Each year, more than 75 000 people are killed on rural roads in OECD Member countries. This represents more than 60% of all road fatalities in OECD countries. The socio-economic costs of the resulting fatalities are on the order of USD 135 billion per year. It is likely that personal injuries in rural road crashes are equally staggering in their number and costs. Unfortunately, data for documenting the rural road safety problem from this perspective is unreliable and inconsistent among the OECD countries. The risk of being killed on rural roads per kilometre driven is generally higher than on urban roads and four to six times higher than on motorways. Rural road accidents are generally more severe than accidents on urban roads due to differences in operating speeds, road geometry, functionality, enforcement levels and other factors. This accounts for the relative share of rural road fatalities in total road crash fatalities which has increased from less than 55% in 1980 to more than 60% in 1996. Because OECD countries have generally experienced a reduction in the total number of road crash fatalities in the same period, it is clear that motorway and urban road safety improvements have been more successful or have been given higher priority than those on rural roads.

The conclusions from these data are inescapable: the rural road safety problem is very serious and all road safety indicators (size, risk, development over time) clearly require priority attention by decision makers and the road safety community. From all appearances, the rural road safety problem has been neglected over the years in comparison to the higher level attention that has been given to the safety problems on motorways and urban/residential roads and streets. This is evidenced by the general lack of explicit safety policies or targets for rural roads in most OECD countries. Given this state of affairs, the rural road safety problem deserves a higher priority in future road safety policies, without, of course, hindering the efforts directed at reducing crashes in urban areas. The OECD therefore created an Expert Group composed of representatives from 13 countries to examine the problems and propose strategies for improving the situation.

Characteristics of the rural road safety problem

Rural roads are defined in this report as roads outside urban areas that are not motorways or unpaved roads. The wide variety of principles and implementation practices used in road classification schemes obscures a correct representation of the size and nature of the road safety problem and makes it difficult to compare rural road safety across countries. In spite of this, it is apparent that as much as 80% of all accidents on rural roads falls into three categories: single vehicle accidents -- especially running off the road, head-on collisions and collisions at intersections.

Single vehicle accidents constitute 35% or more of all fatal rural road accidents. This type of accident is the most prevalent because all three elements of the family of hazard factors -- driver behaviour, vehicle, and road (infrastructure) environment -- contribute to these accidents and increase

9 their severity. Head-on collisions make up nearly 25% of all fatal accidents on rural roads. Driver behaviour and the road environment are the principal factors in these accidents. Collisions at intersections account for about 20% of all fatal rural road accidents. Again, driver behaviour and road infrastructure are the key contributing factors to these types of accidents.

Rural road accidents are scattered over the entire rural road network. Under these circumstances, a pressing challenge for safety professionals is to understand their causes and the contributing factors. A main conclusion from this analysis is that the rural road system itself has inherent characteristics that significantly contribute to the high number of accidents and the high risks.

Inappropriate and excessive speeds are a key factor in rural road accidents because the actual speeds on rural roads are relatively high under circumstances where these high speeds cannot be safely maintained. For example, because of their historical origins, rural roads generally have inconsistent design characteristics over their total length as well as problems in individual design elements. This requires constant speed adaptation to account for regularly changing situations and circumstances, thus increasing the opportunities for human errors and leading to higher risks for accidents. The report therefore concludes that reducing inappropriate and excessive speed together with safe road and roadside design are the key elements to improve rural road safety. Aside from this, fatigue and alcohol/drug use are also key factors in rural safety. Equally importantly, speed variation caused by the presence of buses, heavy trucks, agricultural vehicles, mopeds and bicyclists generates higher accident risks than on other types of roads.

A strategy to improve rural road safety

Rural road safety is completely different than motorway or urban road safety and thus requires a separate management approach. Such an approach is almost non-existent in OECD countries. Knowledge about safe rural road design is developing rapidly, although it is still incomplete. A systematic research approach with generally accepted research methodologies and tools would markedly increase the speed of knowledge development as well as the reliability and usefulness of the results. Specific attention to safety as a basic design element in university-level road engineering courses would also help speed knowledge dissemination. It is therefore recommended that every OECD Member country should develop a rural road safety improvement strategy. It is also recommended that each country should develop short-, short-/medium- and long-term programmes based on a sound analysis of the problems. Such plans must pay special attention to raising awareness of rural road safety both within the general public and within the organisations of all key actors -- i.e. government, peer groups and others.

In the short-term programme, it is advisable to develop and implement a speed management programme in which speed-limit setting and speed enforcement (combined with publicity campaigns) are key components. Also, a trauma management system could be installed in the short term. In the short- and medium-term programmes, traditional infrastructure measures have to be chosen that emphasize investment in the quality of the rural road infrastructure. It is recommended that low-cost, effective and efficient infrastructural measures are selected that preferably fit into existing road maintenance programmes. Targeted “black spot” programmes have proved highly effective in this regard. Long-term programmes should include Intelligent Transport Systems (ITS) applications among other measures.

10 Safety measures

Various safety measures, inlcuding many that are low cost, can improve rural road safety and these are suggested throughout the report. Although a structural network-wide approach is required and recommended in the report, there is a clear understanding that individual low-cost measures can contribute substantially to the safety of the rural road network. For example, individual safety measures that address infrastructure offer the most plentiful opportunities for safety enhancement on rural roads.

The report strongly recommends that safety should receive explicit attention at every level of the process, from the decision to build or rebuild a road to the planning and design stages, through construction and during operation and maintenance. The basis of safe road design is a consistent, hierarchical road network, in which each road category has a particular function to fulfil. Rural roads should therefore be assigned a specific function rather than trying to cater to a varying mix of functions. Also, the design of the road should be consistent with the function and in accordance with the lowest functional use of the road. Several other measures, ranging from straightening horizontal curves to appropriate application of pavement markings and roadside markers, are suggested.

The report stresses the importance of “forgiving” roadside concepts and roadside improvements in general because they can significantly reduce the severity of accidents. There is very high potential for improving overall safety by treating or removing roadside obstacles. Obstacle free zones of between four and ten metres are desirable if the road geometry and right-of-way will allow it. Finally, knowledge transfer and training in the area of roadside safety is a key action area that can contribute to better and more timely treatment of roadside hazards.

Because physically separating opposing traffic is a rather drastic and often impractical approach, the provision of conflict-free overtaking opportunities combined with effective measures to prevent overtaking elsewhere can have many advantages. In addition, for the prevention of certain types of accident, including run-off-the-road, a combination of increasing lane width and shoulder width is an effective approach. In considering intersection collisions, the report concludes that roundabouts are generally the safest solution. However, because roundabouts are a relatively expensive alternative, the decision to install roundabout intersections must be based on a thorough analysis of the cost-effectiveness of this solution in comparison to others. Channelisation as a remedial measure at existing ordinary intersections can be profitable, as can road lighting at intersections to reduce the number of night-time collisions.

In addressing the issue of speed variance on rural roads, separating slow and fast traffic will contribute to the overall safety of rural roads and a number of ways to accomplish separation are suggested in the report. As a final comment on infrastructure, safety impact assessments and safety audits should be undertaken, as appropriate, when planning, designing, (re)building or maintaining roads with the aim to prevent accidents rather than respond to those that have already happened.

Police enforcement is especially important given the contribution of inappropriate and excessive speed to rural road crashes. Effective enforcement can serve as a general deterrent factor that can bring about long-term behaviour changes in drivers if it is coupled with appropriate penalty regimes and publicity. However, due to the great length of the rural road network, enforcement by conventional means has very limited potential. Publicity campaigns associated with targeted enforcement can increase the enforcement effects and contribute to a change in driving norms. In a similar vein, the report concludes that repeated enforcement creates longer halo effects compared to

11 “blitz” campaigns. By introducing a random enforcement element, enforcement effectiveness can also be increased and longer halo effects produced. As well, automated enforcement technologies that target the causes of the principal rural road accidents should be considered. Finally, the report recommends that a portion of the funds generated from traffic enforcement activities be earmarked for rural road safety.

The full potential of ITS solutions for rural road safety can be realised only if research is undertaken that addresses the costs of these systems, specific technical issues, the human-machine interface, and institutional and political constraints. The report identifies a host of low-cost ITS measures that will be ready for deployment within the next three years and that could make contributions in reducing the principal accident types on rural roads. Paramount among these, given the major role of speed in rural road accidents, are speed control technologies such as speed advisory systems and adaptive cruise control. Other near-term, low cost measures include systems for driver monitoring, intersection approach warning and guide lights. Applications such as smart seat belts, air bags and vehicle data recorders will be broadly available and can lessen the rural road safety problem. Decisions to apply higher cost measures in rural road situations must be made on case-by-case basis.

Identifying an accident location is one of the key problems in responding to rural road crashes. The report cites several options that can improve the situation, including: improving road and kilometre/mile identification schemes; expanding the use of GPS; and exploring possibilities for automated accident detection. Available technologies such as cellular telephones are viewed as an extremely positive advance as they can shorten arrival time and improve the overall information available about an accident situation. Publicity campaigns in conjunction with more widespread first aid training can also help to improve trauma treatment at the scene of a rural road accident. The report recommends and describes common guidelines and standard procedures that local hospitals could adopt to improve trauma treatment. In the case of multiple traumas or severe injuries that are likely to surpass the capabilities of the local hospital, the police or others on the scene should be aware of procedures for notifying and obtaining assistance from trauma specialists.

Research needs

There is currently insufficient information available on rural road safety problems to adequately support appropriate policy and investment decisions. This is important because improving rural road safety will require unified methods for collecting and reporting accident data, identifying exposure measures and intervention levels, monitoring and evaluating countermeasures and estimating cost effectiveness and benefit-cost ratios of these countermeasures. With these unified methods in place it is possible to build a sound basis for rational rural road safety policies. Therefore, more systematic evaluation of the effectiveness of countermeasures is necessary based on valid and reliable data. In this regard, benchmarking of rural road strategies may help to improve effectiveness. Finally, further research into rural road safety strategies and several individual rural road safety measures is required to rapidly advance safety improvements on rural roads.

12 Chapter I

INTRODUCTION

I.1 Defining rural roads

Rural roads form an important part of the total road network. As such, they have special features that separate them from other elements of the network. However, it is an illusion to think that all rural roads are the same, because they differ in key characteristics such as function, geometric design, usage and traffic behaviour, road capacities and traffic volumes. They range from motorways in some cases to no more than a paved cart-track in others. These differences can be found within one country and even greater differences can be found in comparisons across countries and are manifested in the road crash statistics of countries in terms of the number of fatal and serious crashes, the accident ratio, accident patterns and their characteristics.1

No formal, accepted international definition exists to classify rural roads. However, it is very clear that a wide variety of rural road types could be distinguished from motorways on one side of the spectrum to unpaved rural roads on the other, and all other types in between.

A rough division between rural motorways and the rest of the rural network could be suggested for the road network outside built-up areas as illustrated in Table I.1. Firstly, a part of the rural road network makes rural regions accessible and connects villages and towns. These are relatively busy roads. If traffic volumes are high enough, these roads are part of the motorway network. This rural network represents possibly 20% of the total road length outside the built-up areas and 80% of the vehicle-kilometres travelled. Secondly, the other 80% of the road network length is composed of country roads with no other function than enabling direct access to properties alongside a road. These roads account for only 20% of the vehicle-kilometres travelled.

Table I.1 Make-up and use of the rural road network

Percentage of network length Percentage of vehicle- kilometres

Primary roads 20 80

Other roads 80 20

1. Throughout the report, the terms “accident” and “crash” are used interchangeably and therefore do not differ in their definitions.

13 For the purpose of this report and the data presented, rural roads include all road types except motorways and unpaved roads. In addition, the description of rural roads in this report does not include roads within villages and towns, but does include entry into villages and towns.

In the same way that rural roads are diversely defined, they can also be classified in various ways. For instance, a functional classification system defines categories such as a flow function for through traffic, a distributor function for distribution within districts and regions and an access function for providing access to destinations alongside roads. As part of these functional classification categories, several countries have defined different types of roads, for example arteries (principal and minor), collector roads and local roads. Other than functional classification, some countries use a classification system based on administrative categories such as national, state/provincial, prefectural or municipal. At times, administrative categories that identify roads as 1st class, 2nd class or 3rd class may be used. In some cases, countries use a combination of functional and administrative categories to classify their roads. Obviously, road classification schemes reflect a wide variety of principles and implementation practices. This hampers and obscures a correct representation of the size and nature of road safety problems and makes it difficult to compare rural road safety across countries.

I.2 Rural road safety

Roads have traditionally been built to facilitate travel from one place to another while allowing traffic to reach its destination as quickly as possible in spite of natural constraints such as rivers and mountains and other obstacles such as privately owned land. In the beginning of this century, the major travel mode in many countries was an animal driven cart which led to a more or less dense network of unpaved, winding tracks connecting towns, villages and settlements. Today, a considerable part of the rural road network still consists of these narrow, winding roads, though they may have been gradually upgraded and paved to cater to a wider variety of transport, ranging from pedestrians, bicycles and agricultural tractors to cars, buses and local delivery vans and trucks. With increasing motorisation, the demand for mobility using high-speed, high-quality roads also increased and resulted in networks of high-capacity roads that often, though not exclusively, fall within the category of motorways.

While optimising mobility is still a major factor in road transport decision making, the high number of traffic accidents and casualties, as well as the high accident risks on particular types of roads, has spurred the increasing inclusion of explicit safety considerations in road design and (re)construction. In recent times, safety in its own right has been considered as a valid reason to improve the road infrastructure though, unfortunately, this is still not as widespread a practice as might be needed to fully address road safety problems. This is especially true for the rural road networks in most countries.

The demand for mobility in rural areas has been greater than within urban areas. This growth is a result of developments such as suburbanisation and rural economic development along main roads. These developments are usually coupled with diminishing public transport options relative to increasing dependence on the use of private vehicles. Also, fewer working hours (shorter working weeks, early retirement), together with higher incomes that allow for higher motorisation have resulted in an increase in social and recreational mobility. Because the time spent travelling is not considered as the most important factor in recreational mobility, a substantial and growing portion of this mobility is on rural roads. Finally, a pressing problem that is growing in many parts of the

14 world is the use of rural roads as shortcuts to avoid congestion on motorways. For road safety this is a problem because risks on rural roads are higher than on motorways. It has been clear to highway professionals for some time that safety on secondary networks -- i.e. any type of rural road other than a motorway -- is not a factor that can be ignored.

I.3 Safety on rural roads is declining

In spite of relatively low traffic volumes, the number of fatal traffic accidents on rural roads has grown to 60% of the total number of fatalities on all roads. In the same manner that the term rural roads encompasses a broad diversity of roads, it should be acknowledged that the nature of the problems that contribute to this poor safety situation, as well as possible solutions for these problems, differ considerably for the various types of rural roads. For example, road reconstruction, low-cost infrastructure measures at high-risk locations or regulations/speed limits with associated traffic enforcement are all possible solutions for rural road problems depending upon the situation. However, while research, development and implementation have taken place on a large scale with a view to improving the safety of the primary network, safety on rural networks has fallen behind relative to urban roads. A combination of factors could explain these high risks:

• many roads are of older design and do not meet the requirements that would be imposed today; • various types of transport occupy the same physical space (two-wheeled vehicles, passenger cars, trucks, buses and tractors); • there are limited right-of-way widths, frequent roadside obstacles and a lack of clear roadside recovery areas; • many roads outside built-up areas permit high driving speeds and have blanket speed limits while the course and design of the roads often requires speed adjustment -- this failure to adjust speed creates a situation in which the severity of accidents is much higher on rural roads; • the detection and location of accidents is sometimes delayed and emergency services take longer to arrive at the scene than for urban roads or motorways; • a priority focus on urban safety improvement has resulted in relative worsening of the rural safety situation. These factors could also be heightened by the increasing traffic volume on rural roads. This is important from a safety standpoint for many reasons. For instance, the growth in the length of the rural road network is generally less than the traffic growth. This implies that the traffic volume on these roads is increasing. Furthermore, the budgets available for construction and maintenance of the rural road network are, in relation to the road length which has to be maintained, diminishing. This means that maintenance of the rural road network will be reduced relative to the increasing traffic levels. This situation leads to a decrease in large-scale maintenance activities and an increase in incidental road maintenance. Consequently, structural safety improvements, as an element of the former category, fade away.

It is not completely clear how these factors impact on rural road safety. On the one hand, it can be reasonably assumed that poor maintenance can lead to dangerous situations, certainly if the road user is not aware of the heightened risk. But, neither can the possibility of lower actual speeds

15 due to poor road conditions be excluded. This could be beneficial to road safety. Likewise, the detrimental effects of diminishing governmental budgets and subsidies for public transport in rural areas are not clear. People living in these areas will become more and more dependent on the use of private cars and, as a consequence, crash exposure could increase. However, another possible outcome could be that more and more “platoons” of vehicles could be created, so that the number of kilometres travelled rises but the exposure for accidents does not increase proportionally.

There are many other factors that can be considered in this light. For example, it is likely that the diversity of vehicles -- i.e. more heavy vehicles combined with more passenger cars, motorbikes, mopeds and pedestrians -- in the same physical situation will increase. Likewise, in some countries, the technology and policy developments in support of larger and heavier commercial vehicles will exacerbate the historical fact that rural roads are not designed for the vehicles that use them. It can reasonably be expected that these two trends will not be beneficial to safety. Considering the many different scenarios and possibilities, however, leads to an inescapable assumption that traffic on the rural road network will increase at a relatively stronger pace than in built-up areas and that this volume increase will likely take place on deteriorating roads. There is, therefore, a strong likelihood that safety on these roads will decrease and that the already discouraging fatality trends on these roads will continue if no actions are taken.

I.4 Safety policy for rural roads

One of the first approaches for solving rural road problems is to apply techniques and methods which have proven effective on higher classes of roads. Unfortunately, this approach does not always adequately correct the problems and, for many countries, has reached its limits of effectiveness. Additionally, treatments used on high volume networks may not be practical on rural roads due to financial or technical constraints. Finally, when rural road safety is buried -- i.e. not explicitly mentioned -- in a national safety programme, it is natural that major urban roads, residential streets and motorways that are highly visible to the public will attract the attention of road safety administrations. This situation suggests that rural roads will not receive the priority attention they deserve. All of these issues make it difficult to appropriately address rural road safety problems from the context of a single set of solutions for the entire road network.

It has therefore become necessary to lay down specific techniques to deal with particular problems on rural roads. These problems include: sharp vertical and horizontal curves; inadequate sight distances for passing, stopping and intersections; narrow lanes and bridges; inadequate signing, markings and delineation; verges; roadside obstacles; a wide variety of users with differing characteristics; and a lack of proper roadside safety hardware. In addition, enforcement and emergency response present an entirely different set of challenges than those presented by the problems listed above. For all of these problems, more safety applications that are tailored to the rural road environment are necessary if further rural road safety improvements are to be realised.

I.5 Objectives of the study

The primary objective of the study was to exchange rural road safety experiences and propose policy and other actions that could be taken by OECD Member countries to improve the situation. To achieve this objective, the following supporting objectives were required:

16 • describe the general rural road safety characteristics in various countries so as to identify and analyse the principal rural road safety problems; • review initiatives to improve rural road safety and compare local network safety management practices in OECD Member countries; • assess potential measures for their capacity to reduce the accident risk in relation to the associated cost.

The report has a target audience of road safety professionals and those managers and policy makers one level below road directors. Where significant policy recommendations are made, the audience is road safety directors and senior policy makers such as Ministers. Summaries of the report are also useful for road safety practitioners.

I.6 Structure of the report

Chapters II and III establish a framework and a foundation to facilitate the examination of the current means for addressing rural road safety situations and the development of new strategic approaches. Chapter II presents a general overview of safety problems on rural roads in the OECD Member countries. It covers topics including general information on the rural road network; the use of rural roads; types of accidents; and accident characteristics and costs. Finally, some national safety goals are presented. Chapter III analyses the factors that contribute to road hazards in rural areas and, in describing these factors, presents the key role that the public authorities play by choosing whether or not to implement certain safety measures. An analysis of rural road accidents is then provided in order to identify the predominant types of accidents that occur and the general causes for these accidents.

Chapters IV, V and VI discuss possible safety improving options for rural roads. Chapter IV summarises the main measures in the road infrastructure domain which have proven to contribute to the safety of rural roads. It gives an overview of the numerous infrastructure measures for improving safety, especially those measures to improve the existing network. The chapter also deals with the relationship between road infrastructure and safety in general and specific road design concepts to improve the safety situation on rural roads. Chapter V discusses the particularly challenging problem of enforcement in the rural road environment. It describes enforcement mechanisms as well as the importance of enforcement of speed limits, intersection behaviour, drink driving, and other specific areas of importance to rural roads, especially the important role of publicity campaigns. The chapter provides examples of enforcement activities and a detailed discussion of enforcement funding, management and priority setting. Chapter VI discusses Intelligent Transport System (ITS) technologies that are currently available and others that are in the development phase that could be used to improve rural road safety. Some issues that are particularly relevant to the use of ITS technologies in a rural setting are addressed.

Chapter VII goes beyond the prevention of rural road accidents to examine the curative measures found in trauma management. The special risks that are prevalent in rural road crashes and related critical issues are covered. The importance of response and treatment timeliness for crash victims is described and ways to improve trauma management in rural areas are suggested.

Chapter VIII builds on the information in the previous chapters to suggest a strategic framework for addressing rural road safety problems in the OECD Member countries. It presents a

17 functional approach for considering the rural road network and makes arguments for rationalising road safety-related policies. A central feature of the chapter is a discussion of the concept of policy integration in the various rural road safety areas. Certain elements critical for achieving policy integration -- i.e. organisational/institutional, financial, and information/data -- are identified and their roles in promoting integration are described.

Chapter IX presents a summary of the conclusions and recommendations from each of the chapters in the report. Among other things, the chapter highlights key rural road safety measures and suggests directions for future national and international rural road safety research.

18 Chapter II

GENERAL CHARACTERISTICS OF RURAL ROAD SAFETY

II.1 Introduction

This chapter presents a general overview of safety problems on rural roads in the OECD Member countries. The information has been gathered from the answers to a short questionnaire that was sent to the national representatives of the countries participating in the Expert Group. In addition, readily available data from other sources have been used and are referenced in the text. The topics covered include: general information on the rural road network; the use of rural roads; the types of accidents; and the accident characteristics and costs. Finally some national safety goals are presented.

It is important to keep in mind how the data are presented throughout this chapter. Box II.1 presents a description of how to measure safety on roads and provides a foundation for the presentations.

Box II.1 How to measure safety

Measuring safety in road traffic is not an easy task. The most obvious indicators of a lack of safety are the occurrence of accidents, the number of people killed or injured or the costs of accidents. However, using accidents as an example, consider that from a statistical standpoint they happen relatively seldom. When examining accident statistics, this leads to a skewed distribution called a Poisson distribution. Within a certain period of time in most places no accidents occur and most people are not involved in an accident. Some places and some people may have one accident within that time period, but only rarely do accidents occur more frequently. This problem can be resolved by using longer time periods for analysis. However, the time periods cannot be too long because a variety of contributing circumstances can change. For example, roads are reconstructed, black spots are treated, people age and driving behaviour changes. Ogden (1996) mentions a period of five years as leading to sufficiently reliable data on accident occurrences.

In order to judge safety (or more correctly risk), it does not suffice to look at accident figures alone. One also has to take into account how many subjects -- i.e. people, vehicles or locations -- are exposed to a certain risk. This means that the number of accidents has to be related to some denominator representing an exposure measure. Common denominators are the size of the population, the number of drivers (sometimes divided by age or gender), the number of vehicles, and the number of kilometres driven under certain circumstances. The number of accidents per exposure ratio is called a rate or risk. Depending on the denominator used, a distinction is made (Trinca et al., 1988) between traffic safety (road-related exposure measure -- e.g. vehicle-kilometers driven) and personal safety (person-related exposure measure -- e.g. population). Hauer (1998) provides a current and comprehensive approach for measuring the effects of safety measures using various means.

19 II.2 The rural road network and its use

Various kinds of data concerning road networks and traffic are available from the OECD International Road Traffic Accident Database (IRTAD) hosted by the German Road Research Institute (BASt). Though data are not always available for all countries, the data are comparable and reliable for all of the reporting countries. Because IRTAD data for 1994 was relatively complete, it was decided to use this year for the current analysis. The results regarding overall road network length and rural road network length are shown in Table II.1. From Table II.1, it is clear that rural roads constitute an important part of the road network.

Table II.1 Length of the urban, rural and motorway networks (in thousand km and %)

Urban roads Rural roads Motorways Total road network COUNTRY Belgium 27.9 113.0 1.7 143 (19.6%) (79.3%) (1.2%) (100%) Czech Republic 57.4 55.6 0.4 113 (50.6%) (49.0%) (0%) (100%) Denmark 20.0 50.5 0.8 71 (28%) (70.9%) (1.1%) (100%) Finland 22.6 75.0 0.4 98 (23.4%) (76.6%) (0%) (100%) Germany (West) 239.0 259.1 9.2 507 (47.1%) (51.1%) (1.8%) (100%) Hungary 55.0 50.4 0.3 106 (52.1%) (47.7%) (0.3%) (100%) Netherlands 52.0 54.9 2.2 109 (47.7%) (50.3%) (2.0%) (100%) Republic of Ireland 3.0 89.3 0.02 92 (3.3%) (96.7%) (0%) (100%) United Kingdom 163.8 222.2 3.3 389 (42.1%) (57.1%) (0.8%) (100%) United States 1 287.9 4 924.3 73.3 6286 (20.5%) (78.3%) (1.2%) (100%) Source: IRTAD, 1994.

Table II.2 Vehicle-kilometres travelled (in million km) and per cent increase on rural roads

Vehicle-kilometres travelled 1980-1995 Country 1980 1986 1995 Percentage increase Austria 22 299 24 503 36 427* 63.4 Denmark 13 167 15 215 19 656 49.3 Finland 17 111 22 520 24 559 43.5 Germany (West) 142 600 161 500 215 100 50.8 Ireland 16 086 17 301 25 338 57.5 Netherlands 29 462 30 893 38 117 29.4 Switzerland 16 620 18 480 22 750 36.9 United Kingdom 108 847 135 197 167 314 53.7 United States 859 113 941 251 1 135 851 32.2 *: 1994 figure.

Source: IRTAD.

20 For a first analysis of the safety situation on rural roads it is important to know both the length of the roads and the intensity of their use. In Table II.2, the vehicle-kilometres travelled on rural roads are shown for several different years. The percentage of kilometres driven on rural roads has grown substantially over the reporting period, assuming only a minor growth in the rural road network length. From the data presented, it appears that the upward trend is likely to continue and potentially worsen the rural road safety situation if one assumes that more travel on these roads will lead to more accidents unless specific safety measures are taken.

II.3 Speed limits and driving speeds

Some countries have clearly shown that speed is a major factor in accidents (ETSC, 1995). For example, in Australia, speed is suggested to be a major factor in as many as 29% of all rural road crashes (FORS, 1996a). Though the impact of speed on rural road safety is still a subject of heated debate, there is no doubt that high speed is a contributing factor to the severity of accidents. Thus, all countries have taken steps to impose speed limits in accordance with the road characteristics. Existing speed limits on rural roads are shown in Table II.3. Some studies have attempted to quantify the link between speed and accident risk. Finch et al., (1994) have reported that a 1 km/h increase in speed results in a 3% increase in injury accidents -- i.e. 1 km/h results in a 5% increase. Nilsson (1981) derived a model that performed before/after studies related to changes in speed. The model showed that as speed increases, the risk of being killed increases to the power of four in relation to the speed change. Likewise, the model showed that as speed increases, the risk of being seriously injured increases by a power of three and the risk of injury increases by a power of two. However, speed is not the only determining parameter. The variation of speed should also be taken into account, particularly for certain types of accidents.

Table II.3 Speed limits by country and type of road in kilometres per hour

COUNTRY Highways, main or national Secondary or regional roads roads Austria 100 100 Belgium 90 90 Canada 90 80 Czech Republic 90 90 Denmark 90 80 Finland 100 80 France 110 90 Germany 100 100 Hungary 80 (motorways)/100 80 semi-motorways Italy 110 90 Japan 60 60 Netherlands 100 80 Portugal 100 90 Republic of Ireland 96.5 60 Spain 100 90 Sweden 110 70 Switzerland 80 80 United Kingdom 96 96 United States 90 (typical) 80 (typical) Source: SARTRE, 1998.

21 The actual speed of vehicles on a road rarely matches the speed limit. This is because the actual travelling speed is influenced by the road function, the length of the journey, the vehicle, the driver and the road characteristics. Also, a speed limit is difficult to set as it must integrate a certain degree of user acceptability and their perception of the credibility of the limit. If a speed limit does not match the expectations of the users and there is no effective speed enforcement, the actual speed will be much higher than the speed limit. For example, in the United States, the average speed (87 km/h) on rural roads is 7 km/h higher than the speed limit. Likewise, in France, the average speed on rural roads (93 km/h) is 3 km higher than the speed limit. In Denmark, the average speed on rural roads is about 89 km/h where the speed limit is no higher than 80 km/h (European Commission, 1997). Finally, in Canada, the average speed on rural roads is 90 km/h, 10 km/h over the limit.

It is crucial to note, however, that average speed does not tell the whole story. For instance, if there is wide dispersion of speed among vehicles on a rural road, there will be a higher frequency of overtaking and other manoeuvres that will increase the risk of accidents. As well, most countries report that a significant percentage of drivers exceed the speed limit, some of them by a wide margin. For instance, in the MASTER Project (Fourth Framework Programme of the European Union) it was reported that: 47% of cars exceed the limit on dual carriageway rural roads in the United Kingdom; as many as 75% of rural road users exceed the speed limit in Sweden; and 5.5% of cars exceed the 90 km/hr speed limit on rural roads in Portugal. As well, speeding by heavy trucks is a special problem. For example, in the United Kingdom it is reported (European Commission, 1997) that 77% of rigid/articulated five-axle trucks exceed the speed limit on single carriageway rural roads and an amazing 92% exceed the limit on dual carriageway rural roads.

Another important element of rural road safety is inappropriate speed (ETSC, 1995). Specifically, though the speed limit set may match the conditions on an average section of road, it can be inappropriate at times because it does not fully account for the risks encountered at specific places along the road. For instance, in places where inadequate sight distances, sharp curves or other rural road hazards present an increased risk, a person operating a vehicle at or below the posted speed limit may actually be travelling at an inappropriate speed for the momentary conditions that they encounter on the road. Appropriate speed is, therefore, a significant issue in relation to rural road safety.

The considerable progress made in motor vehicle secondary safety, the minimisation of injury risks brought about by wearing seat belts, and the overall improvement in infrastructure quality may be leading towards greater tolerance by users of higher speed levels. Because this potentially increases risk, speed restrictions and the “monitoring - penalty” system that must go with them, are still particularly efficient safety measures in most countries. However, speed restrictions and enforcement regimes do not take into account driver behaviour on rural roads during bad weather or similar situations. Specifically, though drivers may lower their speed in these conditions, the change may not be sufficient. Thus, a situation may exist where the average speed on the road is lowered, but the accident risk is increased. This is especially a problem encountered on rural roads.

II.4 Accidents and fatalities

When measuring road safety and the impacts of accidents, there are a number of factors that could be considered to determine the extent of the problem. For instance, temporarily setting aside exposure factors, one could use the number of accidents, the total number injured, the number seriously injured or the number killed. In order to provide a consistent and concise analysis of the rural road safety problem in OECD Member countries, this report will use the number killed (fatalities) as the basic measure for comparison. This is relevant because the enormous differences in

22 registration procedures in various countries makes reasonable comparisons concerning safety on rural roads possible only with respect to fatalities as opposed to other measures such as injuries or property damage. This is a reflection of the serious need for better data -- including exposure data -- which could be generated if IRTAD members were to establish strategic road data requirements that would identify the data that should be collected internationally for comparison purposes. It should also be noted that using fatalities as the sole measure does not provide a complete estimate of the social and economic costs associated with rural road accidents. It does, however, provide a practical and understandable way of examining the problem and its magnitude across many countries.

Official death statistics from various countries may not be directly comparable because the time periods specified for attributing a fatality to an accident are different. For example, in the Canadian province of Quebec, death has to occur within one week of the accident, while all other provinces and territories use a 30-day threshold for an accident fatality. In the IRTAD database, data are comparable because the appropriate corrections have been made. Referring to the Canadian example, data can be included for all of Canada and compared to other countries because it has been corrected. Table II.4 shows the number of fatalities per country on rural roads.

Table II.4 Fatalities on rural roads (1995 except where indicated otherwise)

Number of fatalities Total number Percentage of fatalities Country on rural roads of fatalities occurring on rural roads Australia (1992) 908 1 981 46 Austria 737 1 210 61 Belgium 831 1 449 57 Canada 1 776 3 347 53 Czech Republic 812 1 588 51 Denmark (1994) 326 546 60 Finland 287 441 65 France 5 474 8 891 62 Germany 6 041 9 454 64 Hungary 869 1 589 55 Italy 3 421 7 033 49 Luxembourg (1994) 43 74 58 Netherlands 739 1 334 55 New Zealand 392 581 67 Portugal 1 378 2 711 51 Republic of Ireland 304 437 70 Spain 4 354 5 751 76 Sweden 360 572 63 Switzerland 361 692 52 United Kingdom 2 037 3 765 54 United States 21 246 41 798 51 Source: IRTAD.

To gain a complete picture for OECD countries, it is necessary to add data for which there was no complete IRTAD information available. Table II.5 was therefore developed using data from 1995 on total fatalities from the International Road Federation (IRF) and from IRTAD. To obtain a reasonable estimate of deaths on rural roads from these data, it was conservatively assumed that 50% of all fatalities were on rural roads. From these tables, it can conservatively be estimated that more than 75 000 people are killed on rural roads in the OECD countries on an annual basis.

23 Table II.5 Fatalities on rural roads (1995, except where indicated otherwise)

Number of fatalities Total number Percentage of fatalities Country on rural roads of fatalities occurring on rural roads Greece 1 174 2 349 50 Iceland 12 24 50 Japan 6 335 12 670 50 Korea 5 935 11 871 50 Mexico 2 626 5 252 50 Norway 153 305 50 Poland 3 450 6 900 50 Turkey (1993) 4 197 8 394 50 Source: IRF and IRTAD.

Figure II.1 Indexed occurrence of fatalities on rural roads between 1980 and 1995 (1980=100) *

Austria Belgium Denmark 150 150 150 125 125 125 100 100 100 75 75 75 50 50 50 25 25 25 0 0 0 80 82 84 86 88 90 92 94 80 82 84 86 88 90 92 94 80 82 84 86 88 90 92 94

France Germany Ireland 150 150 150 125 125 125 100 100 100 75 75 75 50 50 50 25 25 25 0 0 0 80 82 84 86 88 90 92 94 80 82 84 86 88 90 92 94 80 82 84 86 88 90 92 94

Italy Netherlands New Zealand 150 150 150 125 125 125 100 100 100 75 75 75 50 50 50 25 25 25 0 0 0 80 82 84 86 88 90 92 94 80 82 84 86 88 90 92 94 80 82 84 86 88 90 92 94

Spain Sweden Switzerland 150 150 150 125 125 125 100 100 100 75 75 75 50 50 50 25 25 25 0 0 0 80 82 84 86 88 90 92 94 80 82 84 86 88 90 92 94 80 82 84 86 88 90 92 94

United Kingdom United States Average for these 14 countries 150 150 150 125 125 125 100 100 100 75 75 75 50 50 50 25 25 25 0 0 0 80 82 84 86 88 90 92 94 80 82 84 86 88 90 92 94 80 82 84 86 88 90 92 94 * Breaks in the lines on some graphs indicate that data was unavailable for the year(s) in question.

Source: IRTAD.

24 Figure II.1 shows that for most countries there has been a general decrease in the number of fatalities on rural roads. However, in considering the information in the figure, the question arises as to whether or not improvements in the traffic safety situation on rural roads has been at the same rate as on other types of roads. In Figure II.2 the average percentage of fatalities on rural roads related to the total number of fatalities for 24 countries are shown. Since the data are not available for all countries for all years from 1970 to 1996, there may be some chance fluctuation in the averages. But it is obvious that since the late 1980s there has been a rather dramatic increase in the relative importance of the rural road problem in relation to urban and motorway fatalities. This change could, in part, be explained by the success of safety measures in urban areas and on motorway facilities.

Figure II.2 Percentage of fatalities occurring on rural roads from 1965 to 1996

65

60

55

50 1965 1970 1975 1980 1985 1990 1995 2000

Source: IRTAD.

Table II.6 shows the fatality risk per thousand million vehicle-kilometres on rural roads. The same risk indicator has also been calculated for urban roads and motorways for comparison.

It is clear that driving on rural roads is much more dangerous than driving on motorways. In most countries, it is even more dangerous than driving in urban areas. The main reasons for motorways to be much safer than the other road types are the physical separation of opposing traffic, a minimum of two lanes to allow safe passing, better intersection design, better overall design standards, wider and flatter obstacle-free roadsides and the access restrictions that apply. On motorways, only motor vehicles with certain characteristics are allowed. On most rural roads no access restrictions apply so that different types of transport take place on the same road. Thus, there are vehicles of very different masses and vulnerabilities moving at various speeds in different directions that must deal with each other in some way. This situation makes the traffic much more complex than on motorways and drivers have to be prepared for a much greater variety of road users and road-user behaviours. Hence, there is a wider variety of accident types on rural roads.

25 Table II.6 Fatality risks per thousand million vehicle kilometres on different types of roads by country (1992, except where indicated otherwise)

COUNTRY Rural roads Urban roads Motorways Austria 24.78 27.29 15.52 Denmark 21.79 12.62 2.52 Japan (1994) 15.17 18.39 6.25 Finland 16.58 11.62 4.82 Germany (West) 21.45 13.07 5.76 Netherlands 18.22 16.58 3.54 Republic of Ireland 13.13 27.79 4.83 Switzerland 20.69 14.06 6.07 United Kingdom 12.27 10.65 3.85 United States 18.52 8.20 5.28 Source: IRTAD.

As it is common to describe accidents by their type, the classifications of these accidents have to be similar in order to be able to make comparisons across countries. But the classifications vary widely. In some countries the categorisation is on the basis of driving manoeuvres, in others by the object of the collision. Mostly a mixture of the two is used. When trying to interrelate these different classifications, it emerges that accidents involving other motor vehicles can be compared relatively well and that collisions involving pedestrians or animals are only partly comparable between countries. Table II.7 shows the most common accident types in four countries as an illustration.

Table II.7 Percentage of fatalities on roads by accident type and country

France Switzerland Hungary Denmark Accident type by (1996) (1992–96) (1996) (1991-95) country Collision: head-on 20.5 16.4 31.0 26.0 Collision: rear-end 6.4 2.6 12.7 9 8.3 27 Collision: 18.0 20.9 intersection/other (only crossing (sideswipe or collisions) angle) Single-vehicle collision 39.5 (including 50.8 23.4 25.0 collisions with animals) Collision with pedestrian 6.2 5.9 21.6 11 Collision with animal (see above) 3.4 0.2 2 (+ other) Source: Individual countries listed.

Although the percentages vary somewhat, it can be seen that the single-vehicle (average of 35%) and head-on (average of 25%) collisions seem to be the most frequent kind of accident with fatal consequences. Non-frontal collisions, commonly at intersections, also account for a

26 considerable proportion (average of 20%) of rural road accidents. These data therefore imply that these three accident types account for as much as 80% of all fatal accidents on rural roads.

Box II.2 Types of accidents and traffic flow: An example from Finland

From an analysis of Finnish accident data by type of road (see Table II.8), it can be inferred that the relative percentage of the different accident types varies depending on the intensity of traffic, i.e. average daily traffic (ADT). The proportion of accidents with vulnerable road users increases as the average daily traffic decreases. On the other hand, the proportion of head-on collisions increases as traffic volume increases. Another result is the reduced risk of run-off the road accidents with increasing average daily traffic.

Table II.8 Percentage of accidents involving personal injury or death by type of road (1992 to 1996)

Dual Two-lane roads carriageway Semi- Main roads Main roads Other Other roads Accident roads (not motorways ADT>6000 ADT<6000 roads ADT<1500 category motorways) ADT>1500 Single vehicle 17.3 35.9 20.8 30.5 30.1 44.6 Intersection 60.3 11.1 38.8 27.3 30.1 15.4 Overtaking 6.4 4.6 3.8 4.2 1.9 1.4 Head-on 1.7 27.8 14.8 13.3 10.5 10.2 Vulnerable user 8.1 3.6 11.9 13.4 20.6 20.7 Animal 0.2 8.5 6.4 8.8 3.6 4.2 Other 6.1 8.5 3.6 2.5 3.0 3.4 Source: Finnish National Road Administration.

II.5 The socio-economic cost

As already mentioned, it is difficult to make a meaningful country-to country comparison of the number of accidents resulting in personal injury or property damage only (PDO). The same is true for cost estimations. International research indicates the following items to be included in estimates of road accident costs: medical costs, costs of lost productive capacity (lost output), costs of property damage, administrative costs and a valuation of lost quality of life (loss of welfare due to accidents). For the first four cost items market prices exist; these can be considered as pure economic costs. This is obviously not the case for the valuation of lost quality of life. For this last cost item, the so-called willingness-to-pay method is considered as the most useful measure as for example was done recently in the European Commission’s Green Paper "Towards Fair and Efficient Pricing in Transport" (European Commission, 1995). The European Transport Safety Council (ETSC, 1997) attempted to calculate the total socio- economic costs associated with reported road crash fatalities, serious and slight injuries. Total costs of 90 billion euros were calculated for all European Union countries. Estimating 50% of this amount to be attributable to rural road accidents, we come to a figure of 45 billion euros in the European Union. Since fatalities in the European Union make up only 37.6% of all rural road fatalities in the OECD Member States, one has to multiply by a factor of 2.66, which gives an estimate of about 120 billion euros for all rural road fatalities and injuries in all OECD countries. The costs of rural road accidents can be estimated to be about 120 billion euros (USD 135 billion) per year in the OECD Member countries and may be much higher.

27 II.6 Safety policy

In the countries surveyed, rural road safety was not specified within the overall national safety targets. General targets regarding the reduction of fatalities and/or injuries and other factors are being set in many countries. A report (OECD, 1994) dealing with targeted road safety programmes showed that general goals are widely used. Not all of these have been quantified. In most countries that reported safety goals, they targeted the number and severity of injuries. A look at the disaggregated goals showed that they concern: • certain groups of traffic participants (young drivers, children, the elderly, pedestrians or bicyclists); • certain age groups (children, the elderly); and/or • certain behaviours (seat belt use, drink-driving, speeding).

Only rarely are programmes targeted towards special accident problems, blackspots or special road sites. Only in Australia (FORS, 1996b) was special mention made of accidents in rural areas. In Finland, a Consultative Committee is assembled every three years which compiles road safety targets (Toivonen, 1998). The last recommendations date from November 1996. Five main areas of activity were chosen:

• to curb traffic growth; • to improve road safety in built-up areas; • to improve interaction among road users; • to reduce the frequency of driving under the influence of alcohol; and • to reduce the number of run-off the road accidents and head-on collisions and to alleviate their consequences.

The last point, especially, is aimed at rural road safety. The intention is to decrease the number of the two most frequent types of accident on rural roads and to reduce the severity of the injuries sustained by:

• improving winter road maintenance to prevent head-on collisions; • raising driver alertness by means of road signs to prevent run-off the road accidents; • “softening” the immediate road environment to reduce the consequences of an impact; and • implementing new types of roads to prevent head-on collisions, for example a centre guard rail plus a passing lane system or a centre guard rail plus a narrow four-lane road. Both types should be safer than two-lane roads and cheaper than motorways.

The United States Department of Transportation developed a strategic plan for the years 1997 to 2002 (USDOT, 1997). The strategic goal is: “To promote public health and safety by working toward the elimination of transportation-related deaths, injuries and property damage.” The goals consist of a reduction in the number of transportation-related deaths and in the number and severity of injuries. In support of this goal, the FHWA aims to reduce the number of highway-related fatalities and injuries by 20% in ten years (FHWA, 1998). Several other specific goals are mentioned,

28 but none of these specifically target rural roads. However, a rural transport safety initiative is being developed by the DOT.

In addition to the above, the American Association of State Highway and Transportation Officials (AASHTO) has adopted a national Strategic Highway Safety Plan that includes 92 strategies in 22 key areas (AASHTO, 1998). The plan includes a target to reduce highway deaths by 5 000 to 7 000 annually until the year 2004. The plan does not specifically target rural roads though many of the recommendations will clearly address rural road problems.

In Denmark the Road Directorate has made proposals for the most important measures to improve rural road safety. They include:

• speed reduction measures; • improvement of the verges along the roads; • measures to improve the perception of curves; • increased distance from carriageway to obstacles along the road; and • measures to avoid overtaking on road stretches where the sight distance is insufficient.

In the French-speaking part of Belgium, a more general approach to rural road safety has been chosen. With the help of a local mobility plan an attempt is made to fulfil the mobility needs of the rural population with an improved public transport system, thus reducing exposure and fatalities.

II.7 Conclusion

More than 75 000 people die every year on rural roads in OECD Member countries. The risk of being involved in a fatal accident on a rural road is generally higher than in urban areas. However, because the rural road network is so large, the accidents and fatalities are scattered over a wide area. This makes the analysis and treatment of rural road safety a highly complex and challenging task. Furthermore, it is evident that accidents on rural roads have a tendency to be more severe -- in terms of the number killed or injured -- than on other roads. As the accidents are more severe, it naturally follows that the costs of these accidents are higher.

Driving on rural roads is accompanied by its own special road safety problems. The risk of getting involved in a fatal accident is almost as high as in urban areas and in some countries even higher than that. The problem is increasing in relative importance. Although the absolute number of fatalities on rural roads -- as on all other types of roads -- is decreasing, the proportion of rural road fatalities in relation to total traffic fatalities has been increasing at a dramatic rate since the late 1980s. This increase has been at almost 1% per year. From 1988 to 1996 the percentage has risen from 54 to 60. It seems that the undeniable progress in road traffic safety that has been made in the last decade was much greater on motorways and in urban areas than on rural roads.

In addition to the above, the following conclusions can be drawn from the information assembled in the present chapter:

• rural roads make up a considerable part of the road network, usually more than 50%;

29 • they are used somewhat less than the other road types. Nevertheless, a high proportion of fatal accidents happen on them -- i.e. high risks exist; • three types of accidents account for nearly 80% of all accidents: single vehicle (run-off), head-on and intersection collisions; and • the comprehensive cost of rural road accidents amounts to more than USD 135 billion (120 billion euros) per year in the OECD Member countries.

A high proportion of rural road use speed -- i.e. drive over-the-limit or at an inappropriate speed. This situation, coupled with the inherent lack of safe road and roadside design in most rural areas, leads to accidents that are generally more severe than those in urban areas or on motorways. In addition to these speed-related challenges, speed variation creates an added element of risk for rural road users. Based on all of these considerations, speed-related factors figure prominently in defining rural road safety.

In view of these results, it is surprising that few countries with road safety programmes target driver behaviour on rural roads. Though there are some emerging rural policies in road safety programmes in some countries, for the vast majority the rural road safety policies are buried among urban, motorway and national safety goals and programmes. Among other items, this is an area that deserves explicit attention in order to reduce the enormous toll of life, pain and money that rural road accidents account for in the OECD Member countries.

30 BIBLIOGRAPHY

ALLENBACH, R. and SCARAMUZZA, G. (1998), “General Characteristics of Rural Road Safety in Switzerland”, personal communication.

BARKER, J, FARMER, S. and NICHOLLS, D. (1998), “Injury Accidents on Rural Single Carriageway Roads, 1994-95: An Analysis of STATS19 (UK) Data”, TRL Report 304, Tranport Research Laboratory, Crowthorne, Berkshire.

EUROPEAN COMMISSION (1995), Towards Fair and Efficient Pricing in Transport. COM(95)691Final, Brussels.

EUROPEAN TRANSPORT SAFETY COUNCIL (1997), Transport Accidents Costs and the Value of Safety, ETSC, Brussels.

EUROPEAN TRANSPORT SAFETY COUNCIL (1995), Reducing Traffic Injuries from Excess and Inappropriate Speed, ETSC, Brussels.

FEDERAL HIGHWAY ADMINISTRATION (1994), Technical Advisory T 7570.2, FHWA, Washington, D.C.

FEDERAL HIGHWAY ADMINISTRATION (1998), “Business Plan 1998-1999”, Safety by Design, Office of Highway Safety, FHWA, Washington, D.C.

FINCH, D., KOMPFNER, P., LOCKWOOD, C. and MAYCOCK, G. (1994), “Speed, Speed Limits and Accidents”, TRL Report 58, Transport Research Laboratory, Crowthorne, Berkshire.

FORS (1996a), “Driving in Unfamiliar Surroundings”, Monograph 5, 1996, Federal Office of Road Safety, Canberra.

FORS (1996b), “Australia’s Rural Road Safety Action Plan”, Monograph 9, 1996, Federal Office of Road Safety, Canberra.

HAUER, E. (1998), Observational Before-After Studies in Road Safety: Estimating the Effect of Highway and Traffic Engineering Measures on Road Safety, Pergamon (Elsevier), New York.

INTERNATIONAL ROAD FEDERATION (1996), World Road Statistics 1991-1995, IRF, Genève.

IRTAD (1997), OECD – International Road Traffic and Accident Database, IRTAD PC User’s Guide, Bundesanstalt für Strassenwesen, Bergisch Gladbach.

31 NILSSON, G. (1981), “The Effects of Speed Limits on Traffic Accidents” in Proceedings of the OECD Symposium on the Effects of Speed Limits on Traffic Accidents and Transport Energy Use, Dublin.

OECD ROAD TRANSPORT RESEARCH PROGRAMME (1994), Targeted Road Safety Programmes, Paris.

OGDEN, K.W (1996), Safer Roads: A Guide to Road Safety Engineering, Aldershot: Avebury Technical.

OSTENSEN, A.G. (1997), “General Characteristics of Rural Road Safety in the United States”, personal communication.

PELCKMANS, J. (1998), “General Characteristics of Rural Road Safety in Belgium”, personal cmmunication.

RUYTERS, H.G. J.C. M., SLOP, M. and WEGMAN, F.C.M. (1994), Safety Effects of Road Design Standards, SWOV, Leidschendam.

SANDERSON, R.W. and ROBBINS, G. (1997), “General Characteristics of Rural Road Safety in Canada”, Personal Communication.

SARTRE (1998), The Attitude and Behaviour of European Car Drivers to Road Safety. SARTRE 2 Reports, Part 1, report on principal results, SWOV, Leidschendam.

SPOUSTA, J. (1997), “General Characteristics of Rural Road Safety in the Czech Republic”, personal communication.

TOIVONEN, S. (1998), Safety Problems of Rural Roads, Finnish National Road Administration (Finnra), Helsinki.

TRINCA, G., JOHNSTON, I., CAMPBELL, B., HAIGHT, F., KNIGHT, P., MACKAY, M., MCLEAN, J. and PETRUCELLI, E. (1988), Reducing Traffic Injury – A Global Challenge, Royal Australasian College of Surgeons, Syndey.

US DEPARTMENT OF TRANSPORTATION (1997), Strategic Plan 1997-2002, Washington, D.C.

WEGMAN, F.C.M. and VAN SCHAGEN, I. (1997), “General Characteristics of Rural Road Safety in the Netherlands”, personal communication.

32 Chapter III

FACTORS CONTRIBUTING TO RURAL ROAD ACCIDENTS

III.1 Introduction

This chapter presents an overview of the nature of accidents in general and then provides a survey of specific factors that can contribute to the three primary accident types -- i.e. single collision or run-off the road; head-on collisions; and intersection collisions. Based on the knowledge of how these factors can individually or collectively contribute to a crash, public authorities can better understand the key role they have to play in improving the safety situation through the measures they do, or do not, implement. Through their action, or lack of action, public authorities can affect:

• road users’ opinions and behaviour; • the condition of the network and its equipment; • the standards for road improvement; • the enforcement; and • the organisation of emergency services. A precise knowledge of accident mechanisms and the many factors involved will provide guidance as to which countermeasures will be most effective in affecting all of the above. This will, in turn, enable roads to be better planned and designed for safety. Much progress still needs to be made in this respect, particularly on low-volume roads, and the chapter provides a good basis for making better safety-related decisions.

III.2 Review of the systemic nature of an accident

As in any approach aimed at improving road safety, the processes leading up to accidents in rural areas need to be examined with a view to identifying the safety problems specific to such areas. Detailed accident analysis methods are now well known. They break crashes down into several phases: the pre-accident situation; the emergency awareness/reaction situation; the shock/impact situation; and the post-accident situation. These methods have demonstrated that an accident results from one or more dysfunctions in a complex system composed of the driver, the vehicle and the road environment (OECD, 1984).

Research has shown that dysfunctions exist such as drink-driving, failure to wear a seat belt, worn out tyres, lack of visibility, speeding and inappropriate speed, and others. An accumulation of these dysfunctions leads to higher accident risks. However, the implementation of effective counter-measures usually requires more in-depth knowledge of all the possible dysfunctions in

33 relation to the environment in which the accidents occur -- e.g. urban or rural roads -- and in relation to the road types -- e.g. motorways, main roads, minor roads.

A significant amount of research has been undertaken to assess the relative importance of each of the three families of hazard factors, which are:

• the driver; • the road environment; and • the vehicle.

Many uncertain elements can affect the results of this research and the various existing road transport regulations or research methodologies can make comparisons difficult from one country to another. But, generally speaking, research results from the United States and the United Kingdom have concluded that 95% of accidents are due to human error -- i.e. observation, decision making and response, 30% result from faults in road design and 10% are the result of mechanical defects (Rumar, 1985).

Whatever the environment, optimum countermeasures must therefore be sought through a series of steps to counteract processes leading to frequent and serious accidents. The steps that can be followed consist of improving driver behaviour, improving vehicle condition and improving road quality. As well, accident alert and emergency call-out services are not always adequately evaluated in road safety research in terms of their impact on the severity of accidents. However, this aspect must not be overlooked in rural areas.

The rest of this chapter identifies the main accident factors related to the driver, road environment and vehicle and re-situates them in the context of rural areas.

III.3 The driver

As pointed out in the previous section, driver behaviour is a predominant factor in nearly all road crashes. This section examines several key items that can contribute to dangerous driver behaviour and thus affect both the likelihood and severity of a crash.

III.3.1 Laws and regulations

A wide variety of engineering, socio-economic and environmental factors generate certain types of road-user behaviour. However, the norms of driver behaviour, or normal behaviour, are set by society through its administrative and legal procedures. In the traffic environment, traffic laws and regulations form the background against which behaviour is judged. To the extent that any behaviour deviates from the norm and contributes to a crash, it is logical to assume that the laws and regulations may have, through their stiffness or laxity, contributed to the situation leading to the crash. In other words, if a law or regulation is not considered realistic or appropriate for a given situation, road users may ignore it and thus introduce risks that otherwise may not have existed.

From another perspective, one must consider how the knowledge of legal and regulatory regimes and their enforcement may affect the behaviour of road users. This knowledge would have an especially significant impact where there is variance between one system of roads -- e.g. motorway

34 network -- and another -- e.g. rural road network. For instance, if someone is in a hurry to get somewhere, if he knows that there is less enforcement on a rural road than on a motorway, this will factor into decisions he makes concerning route choice and the speed with which he operates his vehicle. Decisions such as this can have significant impacts, especially when made in combination with changing weather conditions that can considerably increase crash risks on rural roads that are relatively sensitive to conditions that affect the surface characteristics. In this light, it may be considered that laws and regulations contribute significantly to all three of the main rural crash types.

The effects of speed on safety

As described in Chapter II, speed limit regulations are a special case for rural roads because a large proportion of the public -- sometimes the majority -- does not behave according to the law. The reasons why they do not are, of course, complex and numerous, but include the following: low enforcement levels; improved passive motor vehicle safety; minimisation of impact severity risks produced by wearing seat belts; and the overall improvement in infrastructure and vehicle quality. In a case where the majority of those affected by the law do not obey the law, it could be asked if the law should not be adapted to accord better with “normal” behaviour. However, considering that speed -- i.e. speeding or inappropriate speed -- is a major factor in all three primary crash types on rural roads, it is clear that for reasons of safety the law cannot be adapted to match behaviour without recognising the serious crash risk that will be introduced.

Drinking and driving

Among the behaviour-related hazard factors, alcohol has been the most systematically scrutinised and its impacts on road safety identified. It is clear that alcohol use is a major contributor to all three of the main accident types on rural roads. The evaluation of this factor in accidents varies greatly from one country to another, reflecting both the different regulations and different cultural practices. It is estimated that 0.5 to 3% of drivers take the wheel with a blood alcohol level higher than 0.5 grams per litre, and that this multiplies the accident risk by two. The risk is multiplied by five with 0.7 grams per litre, by 10 with 0.8 grams per litre and by 35 with 1.2 grams per litre. The “alcohol” factor is estimated to be present in some 10% of injury accidents and 40% of fatal accidents in rural areas. However, no definitive study has shown whether this factor is more predominant in, or specific to, rural areas, though a study in Finland showed that the proportion of accidents involving drunk drivers varied considerably depending on road class. On minor roads, the proportion of injuries was found to be about 20% and on main roads it varied from 6 to 10%.

It is useful to note that alcohol tests are more frequent on urban or near-urban roads leading to places of leisure such as discotheques, rather than in the open country where more importance is placed on speed checks. These alcohol tests mainly target young, occasional drinkers as it has been documented (AASHTO, 1998) that young people between the ages of 16 and 24 are disproportionately represented in crashes where alcohol is a factor.

The effects of alcohol on drivers’ vigilance and on their reaction capacities are quite well known. This has led to an interest in other products such as drugs and medicines which are known to be consumed and to induce the same types of risk. Among these products, drugs appear to have a greater impact on urban accidents, as evidenced by reports on the locations where drug-related traffic programmes exist (AASHTO, 1998). The effects of medicines such as tranquillisers must not be underestimated in rural areas, particularly on routes carrying long-distance traffic.

35 III.3.2 The influence of route choice on driver behaviour

Safety studies on rural routes often highlight very different accident typologies depending on whether the people involved are local road users or transiting motorists. A paper from the United States (Blatt, 1998) indicated that three out of four fatal accidents -- i.e. 75% -- on rural roads involved rural and small-town residents. A characteristic driving pattern that could correspond to these motorists is that many of them drive in the “automatic control” mode -- i.e. they know the vicinity too well and so take less care. The other quarter of accidents involves motorists who are unfamiliar with the road and are surprised by a particular situation. On long-distance journeys, other factors, such as the lack of overtaking facilities or a lack of vigilance due particularly to fatigue, may be influential in an accident. Though it is clear that this factor contributes to all accident types, these considerations suggest that route choice, especially in regard to seeking short cuts on rural roads, contribute more to head-on and intersection accidents than they do to run-off the road accidents.

Many theories have been put forward about the driving behaviour of motorists in this context. These theories can serve as a guide for road safety professionals and support decisions on possible countermeasures. Of particular interest, many of the results indicate that the route choice and behaviour of road users can be conditioned by more general determining factors than those detected by accident hazard micro-analysis and the elementary “driver-vehicle-road” system described earlier. Further influential factors are the organisation of road and transport networks, traffic control and regulations on these networks as well as their spatial organisation. For example, rural roads are sometimes chosen as a short cut to avoid congestion on motorways. Additionally, the economic, social and cultural environment that affects the number, purposes, nature and circumstances of journeys will influence driver behaviour.

III.3.3 A driver’s mental representation of the road

Theories on driving behaviour converge on the concept that the motorist reacts to a limited number of visual scenes. This being the case, the driver will try to behave consistently with specific road risks. But, where the situation is unfamiliar, the ability to understand and anticipate is disturbed and can contribute to the initiation of an accident. Moreover, as driver behaviour in this regard is shaped by a host of previous experiences, this might explain the excessive involvement of young people in road accidents, insofar as they are at the learning and adapting stage with little experience to build upon.

An important factor in this regard is that in recent decades, the development of road networks has led to a diversification of the types of roads offered to motorists, both in rural and urban areas. For example, in the 1950s the first motorways came into being. Naturally, they required a long learning phase by the driver to adapt to their new design and operational features. This learning phase is not yet complete in countries that do not have extensive motorway coverage. There are several characteristics that separate motorways from other roads. For instance, they make it easy to drive fast and overtake with virtually no risk of head-on accidents. Grade-separated junctions, central traffic divisions, hard shoulders and a high standard of road equipment are the characteristic features of these roads. Some of these features have subsequently been incorporated into “ordinary” roads such as semi-motorways, express roads or on some major roads in rural areas. Unfortunately, this is a source of confusion for the driver and, as such, may skew his interpretation of what is appropriate behaviour on a given road and therefore increase the crash risk.

36 Similarly, in the past 30 years, there have been considerable transformations on major roads which may have two, three or four lanes with varying road improvement levels. As for minor, or low-volume, roads which are not given any specific treatment, traffic trends have now caused the layout patterns of some of these roads to be modified, thereby making them less easily recognisable. Among other things, this improvement work can result in speeds that are inconsistent with the road configuration. In all cases, such transformations can either present the driver with unexpected situations or lead the driver to misinterpret the nature of the road and therefore increase the risk of an accident. In any event, the misinterpretation of road contributes more heavily to head-on and intersection accidents than it does to run-off the road accidents.

A recent UK study by Clarke et al., (1998) on overtaking accidents set out to analyse 973 overtaking accidents -- of which 410 were studied in depth -- from police records in one county during 1989 and 1993. Detailed causal mechanisms were examined and were found to vary across drivers by age, experience and gender. Two measures in combination -- i.e. not overtaking a vehicle travelling at or near the speed limit and not overtaking in the vicinity of a junction -- would have prevented 54% of the accidents in the sample.

III.4 The road environment

As pointed out in the earlier chapters, the rural road network is not a planned network, but rather an historical network. As such, the road environment, including infrastructure, on most rural networks is a major contributing factor to rural road accidents. The following sections briefly describe some of the critical factors that should be taken into account when considering road safety.

III.4.1 Road legibility and inconsistency

Section III.3 described the difficulty associated with assessing the importance of the hazard factor related to problems generated by a driver’s mental representation of a road. This type of assessment can also be made through macro-analysis that is familiar to most road managers. For example, it is well known that a level intersection on a two-lane dual carriageway with central separation generates particularly severe shearing accidents. Likewise, the construction of a two-lane dual carriageway (no median) with corresponding engineering structures can cause fatal head-on accidents. In a number of studies, these anomalies have been singled out as road legibility deficiencies. To one degree or another, these deficiencies can contribute to all of the principal accident types on rural roads. Thus, in order to complement a driver’s mental representation of the road, the legibility of the road must be improved and its consistency must be ensured.

The roadway geometric design elements -- i.e. vertical and horizontal alignment and cross-section -- are important elements of consistency, especially given the fact that, as stated in Chapter I, rural roads are historical rather than designed roads. Of these elements, the vertical and horizontal alignments control both the sight distance and the safe operating speed of the road (FHWA, 1992a). Thus, a correct and consistent combination of horizontal and vertical alignment promotes uniform vehicle speed and therefore contributes to safety.

37 III.4.2 Surface characteristics

Poor contact in the interface between the vehicle and the road can result in skidding or overturning. This defect can have dramatic consequences, in terms of accident severity, in rural areas where speeds are high. This is especially the case for run-off the and intersection accidents. The problem of poor contact stems from the skid-resistance and/or the evenness of the road surface -- i.e. the surface characteristics.

In regard to deficient skid-resistance, there are three zones that are particularly susceptible to this interface contact problem: 1. Reduced speed areas, particularly intersections, where the longitudinal skid-resistance may be highly stressed. Contact problems in this zone tend to result in rear-end and side impact collisions. 2. Bends, where the transverse skid-resistance may be highly stressed. Problems in this zone usually cause vehicles to run off the road. 3. Transition areas (feeder lanes) where both types of stress (and crashes) are possible.

In all three areas, skid-resistance coefficient requirements need to be very high particularly on heavily trafficked roads -- i.e. more than 10 000 vehicles per day.

As regards road evenness, it seems that some short wavelengths -- between 0.8 and 2.8 m -- can also have an effect in the zones described above, particularly on bends where a bad cross-fall can accentuate the adverse effects of poor skid-resistance. This is especially relevant because accidents on bends probably have the most serious implications for safety in rural areas. This is because bad contact effects are linked to layout problems which are particularly acute on older roads. Isolated bends with a radius of less than 150 metres are shown to have a very high risk regardless of the quality of the road surface. However, any-radius bend will generate risks wherever they are coupled with a surface-related or curve regularity factor that requires the driver to modify his speed.

Another surface irregularity that is relevant to safety on rural roads is wheel path rutting. Wheel path rutting can result from pavement wear or depression caused by heavy loads. Though many elements associated with rutting can increase safety risks, the most serious concern is when the ruts fill with water during rain storms. Water that has pooled in the ruts can cause vehicles to hydroplane -- i.e. lose contact with the road -- and thus lose control. This situation can easily contribute to all three of the major accident types on rural roads.

III.4.3 Roadside obstacles

Loss of control is one of the major factors that can contribute to run-off accidents in rural areas. There are many causes for running off the road, such as: a loss of driving vigilance -- i.e. through a harmless action such as adjusting the car radio, turning to look at a back-seat passenger or speaking on the telephone; poor tyre contact (friction coefficient) with the road; evasive manoeuvres to avoid collisions with animals, pedestrians and other vehicles; poor road legibility leading to drivers being “surprised” by unexpected elements like sharp curves; and overtaking manoeuvres. Many of these actions are often accompanied by inappropriate speed.

38 Run-off the road accidents are all the more serious when the vehicle ends up colliding with a fixed roadside obstacle. The FHWA (1992b) highlights a study in which it was reported that single vehicle accidents per mile per year were highest on rural roads without a clear zone and that rate improved rather dramatically for combinations of gentler side slopes and improved clear-zone policies. In crash situations without clear zones, the impact causes sudden, violent vehicle deceleration. The impact is often inadequately cushioned because it is rarely head-on and causes the vehicle to wrap around the obstacle. This is the case for trees, which account for one-half of all accidents, as well as for power or telegraph poles and signposts. Although considerable progress has been made in vehicle construction to absorb impact when vehicles are involved in head-on collisions, side impact is still a problem area for manufacturers.

It should also be noted, though, that roadside objects are not the only dangerous obstacles. Some design elements such as ditches or banks -- i.e. side slopes or back slopes -- are aggravating factors on the roadside. In some countries there are natural ditches at the roadside that can be quite deep and have steep slopes that require safety barriers but may not have them. Likewise, it is not uncommon in many places to travel on causeways that are two or more metres high and not equipped with safety barriers. In a similar manner, steep side slopes that should be protected with barriers are often left open. In all of these cases, roadside obstacles do not “cause” a run-off the road accident, but they do increase its severity.

III.4.4 Visibility

Sight is one of the driver’s primary sources of information because as much as 90% of the information used by the driver is visual (Lay, 1986). The process that takes place between the time when information is received and a particular driving action is accomplished is complex. Such a process includes a detection phase, an identification phase, a comprehension phase, and a decision phase. Of course, after the decision there will be action and it is only natural that visual and reactive performance varies greatly among different drivers. This is a point that is often overlooked by road planners who use, for instance, a mean visual performance level -- i.e. 10/10 visual acuity -- when designing a road or a road sign. This can partly explain why visibility contributes to crashes, especially head-on and intersection, on rural roads. However, given the situation of ageing populations and the accompanying growth in the number of elderly drivers in most OECD Member countries (OECD, 1998), these criteria will be increasingly under review as road managers and designers try to satisfy a higher number of more diverse motorists.

The problem is the same in relation to the sight distances that are adopted for other road design standards. Though the sight distance at an intersection is calculated by assuming an average time period for a vehicle to turn left from a minor road onto a major road, this time is likely to increase, perhaps significantly, for most elderly people. This change could require the theoretical sight distance to be multiplied by as much as 2 or 3. In a similar manner, the visibility criterion also has a bearing on the design of bends or longitudinal sections in relation to overtaking distances.

It is also important to note that sight distance cannot be isolated from other road characteristics. Specifically, when sight distance is improved, speeds will naturally increase. However, higher speeds could, if no other road improvement actions are taken -- i.e. improved markings, signs, straightening curves, etc. -- lead to higher accident risks. Thus, in undertaking improvements such as this, both positive and negative effects must be considered and how the negative consequences can be countered.

39 Over and above these considerations, there are also, unfortunately, many objects such as road equipment or parked vehicles that can obscure the driver’s visibility. The most common of these objects are signs at road junctions. Such masking elements could easily be eradicated. There are also weather conditions, such as fog (and snow, to a lesser extent) which on rural roads mainly pose problems of visibility. Admittedly these situations are not important factors in regard to the number of accidents, but the crashes that result from them can be highly spectacular and often involve more than one vehicle.

A final comment on visibility has to do with poor visibility at night which can also be a contributing factor in accidents. When it is dark (night time or bad weather conditions) the driver has difficulty perceiving the changes in the road. It is also very common for drivers to be blinded by the lights of oncoming vehicles, especially if the driver has forgotten to switch to low beam. In all of these cases, the distraction of the driver can be serious enough to lead to head-on or run-off the road accidents.

III.4.5 Traffic control devices

The driver depends on receiving visual information through traffic control devices such as signs, signals and pavement markings. Signs and markings fulfil the important function of warning, regulating and guiding road users. They provide extra visual information which supports the existing road design (e.g. roadside and pavement markings), complements it (e.g. local priority regulation signs) or compensates for local design elements which are below standard (e.g. curve or narrow road warning signs). If correct and timely information of this type is not provided to drivers, they will be more likely to make inappropriate decisions and thus be at higher risk of becoming involved in a serious accident. Because rural roads are, by their very nature, dangerous roads -- multiple functions, varying surface characteristics, sometimes limited visibility, sharp curves, etc. -- traffic control devices have a particularly vital role in rural road safety. In addition, the rural road environment may require traffic control devices whose size and placement differ from those on urban roads to account for higher speeds and other characteristics so that the driver can see and react sooner to the information provided. The end result is that poor attention to traffic control devices on rural roads can contribute to all of the major accident categories.

III.4.6 Animals

A special factor on rural roads involves collisions with or accidents -- especially run-off the road -- resulting from wild animals encountered in the roadway. Though they are not a major source of accidents (less than 2%), these accidents are increasingly attracting attention, both because they are constantly increasing in numbers and because they sometimes involve animal species that society would like to protect and preserve. In fact, the growth in the number of these types of accidents can be primarily explained by the growth in the animal population stimulated by protective measures. This situation is a concern on rural roads because collisions with larger mammals in the deer, boar, elk or even kangaroo families can be particularly serious.

Road managers and ecologists tend to know which roads are crossed by wild animals. Thus, warning signs are usually set up in appropriate locations. However, the low probability of encountering these animals undermines the credibility of such road signs with the road user and therefore negates most of their safety benefits. Several possibilities exist for improving the safety situation in this regard. For example, drivers could be better informed and warned through the use of

40 more explicit road signs than those specified in the regulations. As well, verges could be cleared to improve lateral visibility or specific animal under-crossings could be built on some heavily-trafficked roads. Several of the solutions proposed in the following chapters, though not specifically targeting the animal problem, can contribute to reducing the number of crashes resulting from the presence of animals on roads.

III.5 Vehicle/user characteristics

One of the characteristics of rural road environments that separates them from other types of roads is the mix of vehicles and users found in the same space, combined with relatively high speeds. Specifically, unlike motorways where more uniform speeds are found or urban roads where slower speeds are common, rural roads are characterised by mixes of motorised and non-motorised traffic, fast and slow vehicles, and heavy trucks and vulnerable users. All of these blends are dangerous and explain the higher risks to the average rural road user. The following sections present the related problems in detail.

III.5.1 Pedestrians and bicycles

For pedestrians and bicyclists, road safety is often considered as a purely urban problem. However, we cannot ignore these road users and neglect their safety on rural roads because, although these types of accidents are less frequent due to lower traffic levels, they are particularly severe. In Great Britain, 45% of all bicyclist deaths in 1995 occurred on non-built-up roads, although 91% of the total bicyclist casualties occurred on built-up roads -- i.e. urban roads with a speed limit of 64 km/h or less (Gardner, 1998). In the United States (FHWA, 1992c), more than 14% of all non-fatal pedestrian accidents occur in rural areas, but 25% of fatal pedestrian accidents occur in rural areas. A similar phenomenon is reported for bicycle accidents in rural areas -- i.e. 11% of total non-fatal bicycle accidents, but 32% of fatal bicycle accidents. For example, out of 100 pedestrians involved in all accidents on rural roads, 25 are killed. In urban areas, only 5 in 100 are killed. Pedestrian accidents are also differently distributed on rural roads than on urban roads. In urban areas, accidents mainly occur when pedestrians are crossing the road. On rural roads, one-half of accidents happen when pedestrians are walking along the roadside. This is particularly true in some countries where the regulations do not require pedestrians to walk facing the traffic. For bicyclists, one of the few studies that have been made on accident types in rural areas is from Gardner (1998). Speed differential between bicyclists and motor traffic is a key factor, and there is evidence that severity of collisions is directly proportional to speed limits. For instance, the FHWA (1992c) reports that more than half of the fatal bicycle accidents in the United States occur on roads with speeds greater than 35 miles per hour. Other investigations suggest a predominant accident type is probably the rear-end collision by a vehicle, mainly at night. Thus, though pedestrians and bicyclists contribute to the overall number of accidents on rural roads, they do not significantly contribute to the three main accident types.

III.5.2 Slow vehicle traffic

Slow-moving vehicles -- trucks, agricultural vehicles, etc. -- can be disruptive elements for normal vehicle traffic on rural roads by motivating other drivers to sometimes undertake risky manoeuvres, especially passing actions that contribute to head-on collisions. Though slow moving vehicles do not cause many accidents as the risk exposure is low, accidents involving these vehicles are usually very severe because of the great speed differential between a faster vehicle and the slow

41 vehicle or in consideration of two vehicles colliding head-on. The severity of accidents is also higher at times because of the nature of the load -- i.e. hazardous materials -- in the slow-moving vehicle. A related factor is the problem posed by heavy trucks on steep -- i.e. more than 6% -- ascending and descending gradients or on long moderate gradients. In addition to the problem of overtaking by light vehicles in these situations, there is a loss of efficiency in the lorry braking system that can often be combined with lack of experience on the part of the driver. Because these vehicle movements play a part in economic activity -- i.e. agricultural activity for tractors, industrial activity for exceptional transportation -- they accordingly deserve to be mainstreamed into road planning and will continue to be present on rural roads.

III.5.3 Motorcycles and mopeds

Very little is known about accidents involving motorcyclists, even though motorcycle accidents have major implications in certain countries where they account for nearly 10% of fatalities on rural roads. In the United States, for example, 47% of motorcycle-related fatalities occur in rural areas. In terms of exposure to risk (i.e. number of fatalities per one billion kilometres travelled), in the Netherlands, it is four times higher when riding a moped (70.00 fatalities per billion kilometres) and three times higher when riding a motorcycle (54.12 fatalities per billion kilometres) compared to riding a bicycle (17.65 fatalities per billion kilometres).

There are many similarities between motorcycle accidents and light vehicle accidents. However, motorcycle accidents are characterised by specific features, including the involvement of drivers who are often young, inexperienced and with hazardous behaviour tendencies. These factors are clearly linked to the sociological composition of this type of road user and contribute to severe accidents, especially run-off the road and intersection crashes. The problem is made worse when the inexperienced riders have powerful and race-style vehicles. Speed and lower visibility -- in spite of the fact that many countries make it compulsory that the motorcycle headlight be on at all times -- are often factors in these motorcycle accidents.

In regard to road infrastructure, motorcyclists are particularly vulnerable when riding on slippery surfaces or road markings. Motorcyclists tend to avoid some types of pavement grading (chippings) and are generally wary of safety barriers and other aggressive obstacles. For instance, motorcycle riders protest vigorously against the aggressiveness of metal railing supports which are liable to cause decapitation when the motorcyclist skids into them.

Understanding the hazards for motorcycles becomes still more complex in areas where there is an interface with bicycles. In view of the travelling speeds and behaviour patterns of moped riders, the general tendency is to assimilate them with motorcyclists and apply the same types of preventive measures to them. Though this will improve the situation for vulnerable road users, it also creates complications by adding another slow moving -- and vulnerable -- vehicle into the traffic stream. This clearly is not the ideal solution.

III.6 Conclusion

This chapter laid out the underlying general problems and deficiencies that manifest themselves in accidents. In examining the contributing factors, the chapter highlighted the importance of the three families of hazards -- i.e. the driver, the vehicle and the road environment -- and why they should be considered in any proposed countermeasures. The chapter showed that the

42 factors related to driver behaviour and infrastructure play central roles in characterising the rural road safety problem. Equally revealing, the chapter highlighted the unique factors related to the diverse nature of traffic on rural roads -- i.e. the presence of slow-moving vehicles, the operation of motorcycles and the presence of vulnerable users -- and the resulting safety situations that arise from this mix. Finally, the chapter described problems related to animals in the roadway, which is traditionally a rural road safety problem. In all of these instances, the information provides a basis upon which the following chapters can build solutions for the stated problems.

43 BIBLIOGRAPHY

BELANGER, C. (1992), An Overview of Human Characteristics and Limitations Related to Road Safety, University of Toronto, Ontario.

BLATT, J. and FURMAN, S.M. (1998), Residence Location of Drivers Involved in Fatal Crashes, National Highway Traffic Safety Administration, Washington, D.C.

CLARKE, D., WARD, J. and JONES, J (University of Nottingham) (1998), “Overtaking Accidents”, TRL Report 301, Tranport Research Laboratory, Crowthorne, Berkshire.

EUROPEAN TRAFFIC SAFETY COUNCIL (1995), Reducing Speed-related Casualties -- The Role of the European Union, ETSC, Brussels.

EUROPEAN TRAFFIC SAFETY COUNCIL (1995), Reducing Traffic Injuries Resulting from Excess and Inappropriate Speed, ETSC, Brussels.

FHWA (1992a), Safety Effectiveness of Highway Design Features, “Volume II: Alignment”, Federal Highway Administration, Washington, D.C.

FHWA (1992b), Safety Effectiveness of Highway Design Features, “Volume III: Cross Sections”, Federal Highway Administration, Washington, D.C.

FHWA (1992c), Safety Effectiveness of Highway Design Features, “Volume VI: Pedestrians and Bicyclists”, Federal Highway Administration, Washington, D.C.

FHWA (1990), A Users’ Guide to Positive Guidance, Federal Highway Administration, Washington, D.C.

GARDNER, G. and GRAY, S. (1998), “A Preliminary Review of Rural Cycling”, TRL Report 310, Tranport Research Laboratory, Crowthorne, Berkshire.

IRTAD (1994), Under-reporting of Road Traffic Accidents Recorded by the Police at the International Level, Bundesanstalt für Strassenwesen (BASt), Bergisch Gladbach.

LAY, M.G. (1986), Handbook of Road Technology, Gordon and Breach, London.

OECD (1998), Ageing Populations: The Social Policy Implications, OECD, Paris.

OECD ROAD TRANSPORT RESEARCH PROGRAMME (1986), Economic Design of Low-traffic Roads, OECD, Paris.

44 OECD, ROAD TRANSPORT RESEARCH PROGRAMME (1984), Integrated Road Safety Programmes, OECD, Paris.

OECD ROAD TRANSPORT RESEARCH PROGRAMME (1985), Traffic Safety of Elderly Road Users, OECD, Paris.

OECD ROAD RESEARCH (1972), Two-lane Rural Roads, OECD, Paris.

OGDEN, K.W, (1996), Safer Roads: A Guide to Road Safety Engineering, Avebury Technical, Aldershot.

RUMAR, A. (1985), “The Role of Perceptual and Cognitive Filters in Observed Behaviour”, Human Behaviour and Traffic Safety, Evans, L. and Schwing, R.C. (Eds.), Plenum Press, New York.

RUMAR, A. (1997), “Status Reports on Selected European Safety Concepts”, ADAC Symposium.

SETRA-CETUR (1992), Road and Street Safety, Paris.

WEGMAN, F.C.M. (1997), Rural Road Safety in the Netherlands: A Sustainable Safe Road Traffic System, SWOV, Leidschendam.

45

Chapter IV

NETWORK PLANNING AND ROAD INFRASTRUCTURE

IV.1 Introduction

This chapter describes and summarises the main measures in the road infrastructure domain which have proven to contribute to the safety of rural roads. It is not meant to be a detailed guide for practitioners, but rather to give an overview of the extensive variety of possibilities to improve safety by means of, sometimes low-cost, infrastructure measures. As the possibilities for constructing new roads are highly limited for economical and environmental reasons in many countries, the chapter emphasizes measures to improve the existing network. Section IV.2 deals with the relationship between road infrastructure and safety in general and then presents the philosophy, needs and problems associated with the topic. In the subsequent sections, specific road design and infrastructure measures to improve the safety situation on rural roads are discussed.

IV.2 Road traffic engineering and its potential for improving road safety

It is a well-established and accepted fact that human factors are a major contributor to the vast majority of road accidents. However, this does not mean that road safety measures should solely be directed at the road users themselves in order to be effective in reducing accidents. Road-user behaviour does not stand on its own, but is largely dependent on the road environment. Consequently, road infrastructure and design are important means to influence human behaviour and have great potential for increasing safety. For example, based on data from studies in the United States and the United Kingdom, Rumar (1985) reported that the road environment played an important role in about 30% of accidents, the vast majority which had an inter-relationship with road-user behaviour. The number of 30% for road-design-related accidents also emerged from an OECD Workshop organised by the SWOV in 1994 (SWOV, 1994). This type of data suggests that a substantial portion of road accidents could be avoided or mitigated by measures to optimise the road environment. To achieve this, a road design with safety in mind is required. Such a design explicitly takes account of the capabilities and limitations of the road users and involves the following:

• making human error less likely by: − automatically eliciting the desired (i.e. safe) behaviour; and − making undesired (i.e. dangerous) behaviour less likely;

• forgiving to human error, if it occurs, by:

− providing an opportunity to correct an error in time to prevent an accident; and − reducing the consequences once an accident has become inevitable.

47 While safety is understood to be an important criterion in road design today, it is only viewed as an underlying premise in the road engineer’s day-to-day functions. However, safety must be addressed explicitly and not just as part of the underlying job. In a recent report of the Professional Engineers of Ontario (PEO, 1997), three main ingredients are identified which are required to consciously build safety into a road:

• explicit attention to safety; • professional know-how; and • guidance on cost versus safety.

Explicit attention to safety means that safety must be considered in all stages of road and traffic engineering, starting at the level of the decision to build or rebuild a road, followed by the overall planning stage, the design stage, construction, and during operation and maintenance. In other words, the earlier in the process safety is considered, the smaller the need for remedial work afterwards.

A major problem is the professional know-how. The knowledge of the safety effects of road engineering is still very limited, largely because of a lack of systematic research and commonly accepted research tools and methodologies. As concluded in a report by the Transportation Research Board/National Research Council (TRB/NRC, 1987):

“Despite more than one half century of modern road building, knowledge of the safety consequences of highway design decisions is limited. To a large extent the highway community has relied on uncoordinated research without rigorous statistical controls to expand the knowledge about the safety effects of highway design. Conflicting research findings are often left unsolved.”

This has been further corroborated in reports such as Persaud (1992), McGee et al., (1995) and Hauer and Persaud (1996). Equally importantly, professional know-how must begin in the earliest stages of road engineering careers. Specifically, it is essential that university-level road engineering courses incorporate safety as a basic element of the design process. It is essential that such training provide new engineers with a good understanding of safety so that they do not have to rely solely upon on-the-job training before they can include safety in their designs. A 1998 report by Maycock et al., (1998) for the United Kingdom Highway Agency considered the role of driver characteristics in three specific components of the driving task: speed choice, emergency braking and overtaking. The purpose was to provide the highway design engineer with more detailed insights into the types of drivers using the roads and their performances in critical situations.

IV.3 Network planning

A safe road environment starts with the development of a road hierarchy within the network that provides for the various levels of service, from access to mobility. While there are several other characteristics of roadways which define their function, such as traffic volume, design speed, types of vehicles using the roadway and connections within the network, traffic movement and land access are the prime considerations in classifying the roadway. With the proper integration of land use and transportation planning, local roadways primarily provide land access, while through traffic and high operating speeds are discouraged. On the other hand, the roadways at the upper end of the hierarchy, such as arterials and motorways (freeways) are planned to optimise traffic flow and speed, while severely restricting or eliminating all direct access to adjacent lands.

48 Multi-functionality leads to contradictory design requirements and also to higher risks. Table IV.1 indicates the risk levels of different road types in the Netherlands. From this table it is apparent that combinations of multi-functionality, combined use of different transport modes in the same physical space and relatively high speeds and speed differences lead to relatively high risks. It is also clear that the risks on a rural road without any traffic restrictions -- i.e. an “all purpose” road -- are twice as high as on a rural road which is closed to slow-moving vehicles and with a certain level of access control.

Table IV.1 Injury rates in the Netherlands (1986) on different types of roads

Road type Speed limit Mixed traffic Intersecting/oncoming Injury rates traffic per 106 km Residential area 30 yes Yes 0.20 Urban street 50 yes yes 0.75 Urban artery 50/70 yes/no yes 1.33 Rural road 80 yes/no yes 0.64 Express road or road closed to 80 no yes 0.30 slow-moving vehicles Motor road 100 no yes/no 0.11 Motorway 100/120 no no 0.07 Source: SWOV.

Ideally, if each roadway is designed for a specific function with corresponding design characteristics, the clues provided by the road’s design features can elicit the desired driver behaviour, such as travel speed, and draw attention to other road users such as pedestrians and bicyclists. At present, the majority of non-motorway rural roads are multifunctional -- i.e. from inter-urban to local traffic functions -- and accommodate various types of motorised and non-motorised vehicles and pedestrians. This results in a higher accident risk. Discrepancies between function, design and use originate from, among other things, increased motorised mobility, urban development and industrialisation. By (re)categorising the road network and adapting the design to the (new) function while addressing discrepancies between function, design and use, this risk can be reduced. Where 100% mono-functionality cannot be realised, it is best for the design characteristics to reflect the lowest functional category. This concept is expanded in Chapter VIII. An important element of safe network planning is the issue of access control. The elimination of unexpected events and the separation of decision points simplifies the driving task. Access control reduces the variety and changes the spacing of events to which the driver must respond. This results in improved traffic operation and reduced collision experience. As access is one of the primary functions within the road classifications, it is clear that the frequency of collisions increases with the number of access points to the road. “The influence of access control on safety is so strong that the relationship has been established consistently over the years in spite of the difficulties associated with accident studies. One thing is very clear, the most important geometric design element in reducing accidents is access control” (FHWA, 1992). Therefore, from a safety point of view, it must be recommended that direct access to major rural roads be minimised. This can be accomplished, for example, by providing parallel roads which serve a “collector” function for traffic from nearby settlements and properties wanting to enter or leave the main road. Similarly, as stated in the road safety guidelines for Asia and the Pacific Region, no access at all should be permitted at bends, hill crests and at or near intersections (Asian Development Bank, 1996).

49 IV.4 Safety-related road design characteristics

Before turning to the more detailed analysis of individual geometric road design elements, two general design principles which are important for safety need to be mentioned.

Design characteristics need to be consistent with the road’s function and its behavioural requirements.

As mentioned in the section above, a road should be assigned a specific function that requires specific behaviour with respect to road-user interaction and, probably most evident, with respect to speed choice. The design of the road should be consistent with the function and as such elicit the appropriate behaviour. If, for example, low speeds are appropriate on a particular road because it caters to a mixed composition of traffic, the road has to be designed in such a way that road users consider low speeds as appropriate. In other words, as mentioned in the early 1980s by Alexander and Lunenfeld (FHWA, 1981), there is a need for the road environment to act as “positive guidance”, for road users.

Design characteristics need to be consistent along a particular stretch of road.

Road design creates an expectation by road users as to the appropriate behaviour along a particular stretch of road. Since people are known to be relatively slow in adapting to a new situation, inconsistencies in design along the same road may easily lead to inappropriate behaviour. If, for example, a stretch of road which allows for relatively high speeds is interrupted by a sudden narrow curve or a steep descent, this may be handled at too high a speed, even if the change has been announced through warning signals. In principle, it is best to design a road in accordance with the lowest standard on that road. If the road covers a long distance, the transition from one standard to a lower standard (or to another function) should be clearly marked and supported by a transition zone.

The remaining part of this section deals with the main road design elements which are known to affect the safety level of a road or road section. Whereas, for practical purposes, the elements are treated separately, it has to be borne in mind that the ultimate level of safety of a road depends on the consistency of the design in all its aspects.

IV.4.1 Cross section: separating opposing traffic streams

As described in Chapter II, head-on collisions account for a substantial proportion of rural road fatalities. The majority of head-on collisions are the result of cars swerving unintentionally into the opposite driving lane. Only a relatively small proportion of head-on collisions can be attributed to overtaking manoeuvres.

Head-on collisions can be prevented by (physically) separating opposing traffic. This can be realised by including a median space between the opposing traffic lanes. However, for existing rural roads the required extra space cannot be obtained because of restraints such as trees, rivers, adjacent properties, hills and mountains. An easier approach on existing roads -- although a rather drastic measure -- is physical separation by means of the application of narrow steel, concrete or cable barriers. Whereas it has been reported that steel and concrete barriers could themselves cause accidents, the more serious, often fatal, head-on accidents are virtually eliminated. Based on a meta-analysis of 32 relevant studies, Elvik (1995a) estimated that the installation of a median barrier would result in an increase of the overall accident rate (all severities) of around 30%, but would bring

50 about a decrease of 32% in the probability that an accident would produce fatal injury. A majority of the studies on this subject deal with motorways and no hard evidence for rural roads is available.

Experiences with cable barriers as a median separator are still limited, though a first experimental site has been installed in Sweden in the framework of the so-called “Zero-vision” strategy for road safety. Experiences on motorways in Denmark showed that running into cable barriers may lead to very high deceleration rates and hence severe injuries, despite their large deflection capabilities. Their large deflection capabilities would also require sufficient space, which makes them less practicable for use on ordinary rural roads. In principle, every road with a flow function, independent of its traffic volume, should benefit from separating opposing traffic streams by using a safe physical direction-separating measure. However, in practice, it is the medium- to high- volume roads that are treated in this way. More research is needed on the safe design of these physical measures on two-lane rural roads and on comparing physical means with other “softer” measures using markings only.

One way to reduce head-on collisions caused by overtaking manoeuvres is the provision of conflict-free overtaking opportunities. This is particularly true for roads where the possibility to overtake is limited due to insufficient sight distances (e.g. in hilly areas) or because of high traffic volumes in the opposite direction. A number of studies report a decrease in the number of accidents after the construction of regular overtaking lanes that are open to traffic in one direction only. The effects or from this measure are felt both downstream and upstream if the overtaking lanes are signalled in advance. For example, a study for Transport Canada (ADI, 1989) concluded that overtaking and climbing lane installations have been shown to decrease accident rates by 25% compared to untreated two-lane sections. After reviewing several studies, Elvik et al., (1997) reported that climbing lanes reduce injury accidents by 20 to 40%. As a type of extension of the overtaking lane idea, three-lane roads (2+1), with the middle lane alternately assigned to one direction, must be mentioned. Again, experiences in terms of accident reduction have been positive. In Germany, the accident rate (accidents per 106 vehicle-km) on 2+1 roads was found to be 0.33 compared to 0.48 on two-lane roads, with both road types having a similar total width of 14 m (Brannolte et al., 1992). In Finland, on the other hand, overtaking lanes seem to have a much smaller effect on injury accident reduction, and may be even close to zero (Pajunen and Kulmala, 1988).

IV.4.2 Cross section: increasing lane and shoulder width

In general, accident statistics show that the level of safety improves as the road width increases. For example, Zegeer and Council (1992) reported on a cross-sectional study in the United States that, by isolating the effects of lane width from other road and traffic features, predicted the following accident reduction effects for related accident types (run-off, head-on, opposite and same direction sideswipes):

Lane widening Reduction in related accidents 1 ft (0.3 m) 12% 2 ft (0.6 m) 23% 3 ft (0.9 m) 32% 4 ft (1.2 m) 40%

These results are reported to be valid for two-lane rural roads with lane widths between 2.4 m and 3.7 m and daily traffic volumes between 100 and 10 000. On the other hand, neither should lanes be too wide as this may encourage higher speed and, hence, lead to more accidents or more

51 severe accidents. However, a French study by Gambard and Louah (SETRA/CETUR, 1992) showed that the effect of lane widening on speed is larger on narrow roads (lanes smaller than 3 m) than on wide roads (lanes equal to or wider than 3 m). Obviously, all aspects have to be carefully balanced in order to decide on the optimal lane width.

The benefits of widening shoulders are even less clear. In general, it can be concluded that widening shoulders is less effective than widening lanes, but several studies show (Ogden, 1996) that adding (hard) shoulders to roads where none existed does result in fewer accidents. As well, it appears that a combination of increasing lane width and shoulder width is most effective. For example, when a road with 2.7 m lanes and no shoulders is widened to 3.7 m lanes and 1.8 m shoulders, accidents were reduced by around 60% (TRB, 1987). Table IV.2 presents recommendations based on extensive analysis of existing information and cost-effectiveness calculations by the United States Transportation Research Board (TRB) concerning lane and shoulder widths for roads with speeds higher than 50 km/h. Other studies (FHWA, 1992) have resulted in similar findings.

Table IV.2 Recommended lane and shoulder width in the United States

Traffic flow > 10% trucks < 10% trucks (ADT) Lane width (m) Lane + shoulder Lane width (m) Lane + shoulder width (m) width (m) 1 - 750 3.0 3.7 3.0 3.7 750 – 2 000 3.7 4.6 3.3 4.3 > 2 000 3.7 5.5 3.3 5.2 Source: TRB, 1987.

IV.4.3 Cross section: separating slow- and fast-moving traffic

Large differences in travel speed make the traffic situation less predictable and lead to more braking and other evasive actions as well as more overtaking manoeuvres. On rural roads, large speed differences occur if the road is used jointly by motor vehicles, agricultural vehicles, bicyclists and pedestrians. Thus, efforts to separate slow and fast traffic will contribute to the overall safety of rural roads. There are a number of ways to realise at least some degree of separation, including: i) a parallel road or secondary traffic area for all types of slow-moving vehicles; ii) a parallel, physically separated bicycle/pedestrian lane; iii) a lane at the outer side of the normal running lane for bicycle/pedestrian use only; and iv) a multi-purpose lane at the outer side of the road which is, in principle, assigned for bicyclists/pedestrians, but which may be used by slower-moving motor vehicles to allow faster traffic to overtake. All alternatives offer a safety benefit compared to a non-separation situation, though the largest benefit will be realised if all slow-moving road users are physically separated from the main running lanes (secondary traffic areas) and the smallest benefit will result if there is occasional interaction between slow- and fast-moving traffic (multi-purpose lanes). In fact, multi-purpose lanes are banned or discouraged for safety reasons in a number of countries.

IV.4.4 Cross section: bridges

Rural highway bridges are sometimes associated with accident problems due to problems such as narrow width, poor sight distance and poor signalling or delineation. However, the features that are most important in terms of affecting the accident rate are the bridge width and/or the width of

52 the bridge in relation to the approach width (FHWA, 1992). For consistency, whenever possible the cross section of the bridge in terms of lane and shoulder widths should match that of the roadway. Based on work by Turner (1984) related to bridges on two-lane roads, it was found that the number of accidents per million vehicles decreases as the relative bridge width -- i.e. the bridge width minus the width of the travelled way -- increases. It was found that shoulders of 1 meter or more should be provided on each side of the bridge.

Because of the high cost of bridge replacement, it is often not feasible to increase the width of bridges in rural areas. Under these circumstances there are steps that can be taken to improve the safety of existing bridges to counter the ill effects of poor cross section. Specifically, the following three primary areas have been suggested for attention when considering safety improvements on bridges (FHWA, 1998):

• Bridge -- Improvements to elements of the bridge structure or improvements that help mitigate potentially hazardous bridge features. The following items should be considered: bridge rails, sidewalks, open joints, deck surface (friction surface or potholes), snag points on the bridge rail or bridge abutments, delineation of narrow bridges and strengthening of truss members. • Approach -- Improvements to approach guard-rail alignment, transition sections between guard-rail stiffness and rigid bridge rail, connections between the guard-rail and the bridge rail, drainage features, curbs and sidewalks. • Operational -- Improvements to access points, signs, delineation and pavement markings in the bridge approach area and on the bridge.

IV.4.5 Intersections: type, location and some design aspects

Approximately 20% of rural road accidents happen at intersections. Grade-separated intersections are the safest type of intersection. They are mandatory for motorways, but are also applied at some types of high-volume rural roads. For the majority of rural roads, however, constructing grade-separated intersections for the purpose of safety only may not be cost-effective. T-grade intersections exist in many configurations. Three-way intersections and staggered intersections are to be preferred over the traditional four-way cross intersections. Roundabouts are the safest t-grade intersection type. Barton (1989, cited by Ogden, 1996) calculated the mean casualty accident rate (per 107 entering vehicles) for different types of intersections in Australia. For rural roads, Barton calculated the following accident rates:

Cross intersection, unsignalised: 5.2%

T-intersection, unsignalised: 3.3%

Staggered-t 2.9%

Roundabouts 1.6%

In Denmark, conversion of nine four-way intersections in rural areas into roundabouts with an average island diameter of 25 m resulted in an 83% reduction in accidents and a 90% reduction in the number of injuries, correcting for the general accident trend (Danish Road Directorate, 1994). Schoon (1993) found a 42% reduction in accidents (including damage-only) and an 86% reduction in

53 the number of injuries for rural area roundabouts. The very good safety record of roundabouts in comparison to three-way and, in particular, in comparison to four-way intersections is also reported for the European situation in general by ERSF (1996). The explanation given for this good safety performance is the decrease in the number of conflicts, the low speeds and the small conflict angles at roundabouts. The cost of the additional land required to construct medium- or large-scale roundabouts may affect the cost-effectiveness of these intersections in some countries.

An intersection should be located such that it is clearly visible from all approach directions. Advance direction and warning signs as well as landscape features -- e.g. planting -- can help drivers to better judge the situation they will encounter at the intersection. From this standpoint, intersections located just after a curve or a hill must be avoided. Furthermore, the intersection sight distance (the sight triangle) must be large enough to allow drivers to see any approaching traffic on the other legs of the intersection and to judge their approach speed. With this information, drivers should then be able to cross the intersection safely when starting from a waiting position. According to ERSF (1996), the basic design safety principles for intersections are that they be:

• conspicuous from all approaches to allow safe speed adaptation and lane choice; • meet sight requirements to enable safe crossing from a waiting position; • simple and understandable; • driveable; • if possible, designed to reduce speeds and conflicting angles between vehicle paths; and • enable mutual co-ordination between drivers and vulnerable road users.

Channelisation as a remedial measure at existing ordinary intersections has been found to have a substantial effect on the number of injury accidents. In the Norwegian Road Safety Handbook (Elvik et al., 1997) the following data is presented:

Type of channelisation Change in injury accidents

Full channelisation in four-way junctions - 27%

Channelisation of side roads in four-way junction - 17%

Left-turn lane on main road in four-way junction - 4%

Channelisation of side road in T-junction +18%

Left-turn lane on main road in T-junction - 27%

The report concludes that channelisation can be profitable even at intersections with an average annual daily traffic (AADT) of less than 7 500 vehicles.

IV.4.6 Intersection: railway crossings

Railway crossings are intersections that differ from other intersections only because they are the intersection of a railway with the roadway rather than the intersection of two roadways. While collisions at railway crossings represent only a small proportion of the total road-related collisions, they tend to be more severe given the difference in size between the train and the motor vehicle.

54 However, most collisions at railway crossings do not involve a train, but are collisions between vehicles or a vehicle colliding with crossing infrastructure (Ogden, 1996). In a recent study for the Transportation Association of Canada (Hauer and Persaud, 1996), it was estimated that collisions involving trains account for only about one-third of all collisions at railway crossings.

Since, as mentioned, railway crossings are another type of intersection, the road and railway should intersect as near as possible at right angles to provide maximum sight distance along the track in both directions from both approaches. Similarly, the intersection should not be located on a horizontal curve of either the road or railway. The crossing width should be as wide as the approach road and shoulders. Vertical alignment should be as flat as possible to enhance the drivers' sight lines of the railway, and also to avoid situations where braking distances may be affected. In the case of long heavy vehicles, a level crossing avoids the possibility of the truck grounding out on the crossing.

IV.4.7 Alignment: horizontal curves

A relatively high number of accidents occur on horizontal curves as compared to tangent sections of the road. In particular, run-off the road accidents are more common on curves than on the adjacent tangent sections. Head-on accidents, which are the most serious accidents in terms of personal injury, are also more common on these curves. Persaud (1992) found that in Ontario, Canada, 3.5% of all non-intersection accidents on two-lane rural roads occur on horizontal curves. Taking only fatal accidents into consideration, this figure becomes 21%. Comparing data from Canada and the United States, this study also reported that the average accident rate for horizontal curves is about three times that of tangent sections, and the average single vehicle run-off the road accident rate for curves was about four times that of tangents. On two-lane national roads in France, 25-30% of non-intersection injury accidents occur on curves while on secondary roads it is closer to 45% (SETRA/CETUR, 1992).

The most dangerous curves are isolated curves and the first in a series of curves after a straight or almost-straight section. There are two main, interrelated causes that can be identified as follows: • the curve is negotiated at inappropriate speed, resulting in loss of control of the vehicle which, as a result, enters either the roadside or the other lane; • there is limited sight distance when approaching the curve and in the curve itself, causing an incomplete assessment of the static and dynamic situation which may result in inappropriate speed and position choice.

There are a number of general safety principles related to design, location, pavement and roadside characteristics of curves which can help to reduce these problems. For example, the design elements of a particular road need to be consistent in order not to violate drivers’ expectations and to allow for correct anticipation. A series of relatively wide curves should not be followed by a very narrow one without extensive warning and/or physical speed-reducing measures. Furthermore, it must be possible to negotiate an isolated curve or the first in a series of curves at a speed which is not excessively below the speed which is maintained on the straight section preceding it. Whereas there is a general trend for accident rates to increase as a curve becomes narrower, from a safety point of view the consistency between curves along the road is at least as important.

Another principle to bear in mind is the importance of being able to distinguish clearly between situations where overtaking is or is not possible. To achieve this, intermediate curve radii

55 must be avoided between approximately 800 and 2 000 m. Also, the distance between two successive curves or between a straight section and a curve should be long enough to allow drivers to judge and interpret the situation -- the general recommendation is approximately three seconds driving time. Similarly the sight distance both in the approach to a curve and in the curve itself needs to be sufficiently long for drivers to make the correct decisions concerning speed and position. Curves at the top of a (steep) hill should therefore be avoided.

As shown in Table IV.3 (FHWA, 1992; Elvik et al., 1997), straightening horizontal curves is an effective accident-reduction measure. Note that degree of curve, as used in the United States, can be converted to radius of curve, which is more common in Europe, as follows: Radius of curve equals 1746.3/Degree of curve.

Though these are promising statistics, reconstructing existing curves is expensive and probably only cost-effective on higher-volume roads. There are a number of less expensive measures which may help to reduce accident rates in curves such as the removal or protection of roadside hazards, flattening side slopes, improving skid resistance, increasing the superelevation, paving the shoulders and eliminating pavement edge drops (Zegeer et al., 1990; Ogden, 1996). Typical low-cost measures include upgrading the pavement edge line and centre line, adding raised reflective pavement markers or upgrading the advance warning -- e.g. by speed-activated warning signs.

Table IV.3 Accident reduction in the United States and Norway as a result of flattening curves

FHWA

Original degree of curve New degree of curve Reduction in number of accidents

30 25 16%

30 20 33%

30 15 50%

30 10 66%

30 5 83%

Elvik

Original curve radius New curve radius Reduction in number of accidents

< 200 m 200-400 m 50%

200-400 m 400-600 m 33%

400-600 m 600-1 000 m 23%

600-1 000 m 1 000-2 000 m 18%

1 000-2 000 m > 2 000 m 12%

Source: FHWA, 1992 and Elvik et al., 1997.

56 IV.4.8 Alignment: gradients and vertical curves

Both gradients and vertical curves affect the overall safety level of a road. FHWA (1992) reported that downgrades have accident rates that are 63% higher than upgrades. Gradients are particularly important if the road carries many heavy vehicles. In this case, long (>1 km), steep (>4%) gradients should be avoided (ERSF, 1996). In general, 6 to 7% gradients must be used with care (SETRA/CETUR, 1992). When going downgrade, excessive speed and inappropriate braking -- especially by heavy vehicles -- can easily lead to loss of control. When going upgrade, sight distance is limited by the top of the hill. This is known as the blind effect of a convex curve. Because of the limited sight distance, overtaking is usually not safe and, hence, usually not allowed, on a two-lane road in an up-hill situation. The limited sight distance also means that the situation just over the top is not sufficiently visible to allow a vehicle to stop -- i.e. stopping sight distance -- in time if necessary. Speed-reducing measures may be necessary to guarantee sufficient stopping distance in these locations. However, because modification of the vertical alignment of existing roads is very expensive, more research into the effect of the different vertical design elements on accident occurrence is necessary in order to determine the cost effectiveness of different measures.

IV.4.9 Transition zones between rural areas and built-up areas: gateways

Roads that serve a through function in small villages often lead to severe safety problems because it is common that speed is insufficiently reduced when entering the built-up area. This type of safety problem can be avoided by building by-passes and, as such, diverting through traffic away from villages altogether. This is an effective solution, but its relative expense often makes it infeasible. Therefore, much work has been done to find alternative means to ensure that car drivers adapt their behaviour, in particular their speed, on the approaches to towns and villages.

In order to modify speed behaviour, a number of measures are necessary which, at the village entry, will culminate in what is called the “gateway” to the village. The gateway principle is based on the fact that the perception of speed is influenced by the visual information available to the driver. An important aspect that affects subjective speed perception is the width of the road relative to the height of the surrounding vertical elements. Drivers tend to reduce their speed if the vertical elements along the route are higher than the width of the road elements (ETSC, 1995). This can be achieved at relatively low cost by road narrowings, possibly supported by advanced progressive speed limit reduction, rumble strips and/or pavement markers in combination with (properly protected) high trees, light poles or buildings alongside the road. The gateway should coincide with the start of the local speed limit and should comprise (physical) speed-reducing measures. However, the creation of a gateway without any physical speed-reducing measures is much less effective. Furthermore, it is important to complement the speed-reducing measures at the entry of a village by measures along the through road within the village in order to maintain the speed reduction.

In the United Kingdom, a number of trials of schemes to reduce traffic speeds on the approaches to and through rural villages have been tried with some success. The VISP Project completed in 1994 (TRL, 1994) covered 24 villages, and subsequently there have been schemes on main roads through towns and villages to reduce the danger of fast-moving traffic using various gateway, countdown and traffic-calming schemes (TRL, 1996; Barker et al., 1997).

57 IV.5 Roadside safety

Roadside design is in fact part of the cross-section design of roads and should, as such, have been treated in the previous section. There are two compelling reasons for dealing with this issue in a separate section. First, there is wide variation in regard to the inclusion of roadside design as part of the common practices for road design. Secondly, the potential effect of treating the roadside on the overall safety level is very high, in particular for rural roads, and hence needs explicit attention. Thus, this section covers the subject in some detail.

IV.5.1 The size of the problem

As reported in Chapter II, approximately one-third of the fatal accidents on non-motorway rural roads can be classified as run-off-the-road accidents (also see Slop and Catshoek, 1995; Sanderson, 1996; NCHRP, 1997). The number of these accidents can be reduced by providing for wider running lanes, (hard) shoulders and flat verges so that there is a better opportunity to recover from loss of control (see Section IV.4). If recovery is not successful, the consequences of a run- off the road accident range from fatal personal damage to minor material damage. An important factor that determines the final outcome of these accidents is the direct surroundings of the road -- i.e. the risk of hitting a tree, rock, utility pole or the presence of steep side slopes and ditches. The severity of collisions with roadside obstacles varies with the type of obstacle. SETRA/CETUR (1992) reports the highest mortality rate for tree accidents (25.1 fatalities per 100 accidents) followed by utility pole collisions (17.1 fatalities/100 accidents). Guard rails were found to have the lowest mortality rate (9.7 fatalities/100 accidents). Tree accidents account for more than 25% of all obstacle accidents in France and are therefore the most frequent type of accident.

IV.5.2 Creating obstacle-free roadside zones

One way to reduce the number of collisions with roadside obstacles is to create a sufficiently wide obstacle-free zone along the road by removing the injury-inducing obstacles. This increases the chance for recovery when a car leaves the roadway and reduces the consequences if recovery actions are not successful. Roadside obstacles consist not only of trees, ditches, rocks, utility poles and the like, but steep slopes (upgrade or downgrade) must also be considered as obstacles. The benefits of clearing these objects from the roadside were shown in France where, on the basis of an accident-prediction model, it was estimated that the current annual 1 500 fatalities on major rural roads could be brought down to 500 if all of these roads had a 4 m obstacle-free roadside zone or a continuous guard rail.

The recommended values for the minimum obstacle-free zone differ largely between countries and are based on a very limited amount of empirical data. In some countries, the minimum value is fixed and in other countries the minimum value is made dependent on road category, speed and/or traffic volume. For example, the French design standards for main rural roads recommend an obstacle-free zone of 4 m for existing roads and of 7 m for new roads or when placing new obstacles on existing roads (SETRA, 1994). The Dutch guidelines recommend an obstacle-free zone of 10 m for motorways, 6 m for main single- or dual-carriageway rural roads (design speed 100 km/h) and 4.5 m for single-carriageway rural roads (design speed 80 km/h) (RONA, 1987). Based on current practice in EU Member States and on available research data, Schoon (1994) comes to a “best practice value” for obstacle-free zones of at least 5 m for rural non-motorway divided roads and rural undivided roads with a design speed of 100 km/h, and 3.5 m for undivided primary rural roads with a

58 design speed of 80 km/h. As an example, Box IV.1 shows the experiences in Denmark with a pilot project on obstacle-free zones in rural areas.

In Sweden, more than 25% of motorists killed on the roads collide with a fixed object on the roadside. Half of these fatalities occur in collisions with trees. The design of the roadside area and the various fixed objects located there largely determines the consequences of an accident. The greatest risk of being injured or killed by fixed objects occurs in collisions with trees or culverts. There is a similar risk on collisions with poles, but this is lower than with trees or culverts. Table IV.4 shows the relative accident costs for hitting a light pole in different locations and at various speed limits.

Based on these findings, Nilsson and Wenäll recommend the following measures that can be taken to reduce lighting pole collisions and accident consequences while retaining lighting:

• all newly built environments with lighting should be provided with yielding poles and light poles should always be placed on the inside of curves; • replace steel-tube poles with yielding poles; • replace poles with fewer light sources by longer poles with several light sources; • replace poles with suspended lighting where there are suitable buildings for this purpose; • replace twin lighting poles by lighting poles in the central reservation if such exists.

Box IV.1 Impact analysis of obstacle-free zones in Denmark

In Denmark, a new set of guidelines were developed for roadside obstacles. The new guidelines were then tested on a road section in Fredriksborg County. The following information describes the test road:

• 14 km state road section; • Speed limit: 80 km/h; • Average Annual Daily Trafic (AADT): 6 500 - 8 000; • Accidents (five years): 10 accidents, 8 injuries, 1 killed.

The project involved 50 single-spot improvements at a total cost of USD 85 000. The spot improvements consisted of treating -- i.e. moving or removing, protecting with guard rails or softening -- obstacles within 9 metres of the road side. A subsequent accident study was carried out and showed that the assessed impact over five years would be a reduction of 1.5 accidents with two injuries. A first year rate of return of 53% was calculated and showed that roadside improvement activities are very feasible.

Table IV.4 Relative accident costs for hitting a light pole

Speed limit Straight Straight Inside Outside Junction Junction section section curve side curve side centre side centre side 50 km/hr 1.34 2.55 1.00 3.35 1.79 1.36 70 km/hr 2.67 5.07 1.74 6.67 3.11 2.36 90 km/hr 7.15 13.58 2.84 17.87 5.07 3.85 Source: Nilsson, G. and J. Wenäll (1997).

59 One particular problem in relation to roadside design is drainage structures or ditches. It is generally the case that ditches are not considered as part of the design process, but are left, instead, to hydrology engineers who have the principal aim of efficiently dispersing water. Unfortunately, hydrology engineers are often unaware of the safety consequences of their work. Given the results of a recent analysis of Canadian fatal accident data (Sanderson, 1996) which showed that 22% of the single vehicle run off the road accidents involve hitting a ditch or embankment (most likely the backslope of the ditch), this is an issue which needs urgent attention.

The design of the sideslope -- also referred to as foreslope -- has the potential to either increase the severity of an accident or to help prevent it. Specifically, flatter sideslopes can reduce the incidence of rollover accidents. Approximately 55% of run-off-the-road rollover accidents result in injury and 1 to 3% result in a fatality (FHWA, 1992). Only pedestrian and head-on crashes result in higher injury percentages. Table IV.5 shows the percentage reductions of single vehicle and total accidents as a result of sideslope flattening. For example, flattening a 2:1 sideslope to 6:1 should result in a reduction of approximately 21% of single vehicle accidents and 12% of total accidents. The FHWA report concludes that sideslopes of 5:1 or flatter are needed to significantly reduce the incidence of rollover accidents.

Table IV.5 The percentage reductions of single vehicle and total accidents as a result of sideslope flattening

Sideslope in after condition Sideslope 4:1 5:1 6:1 7:1 in Single Total Single Total Single Total Single Total "before" vehicle vehicle vehicle vehicle condition 2:1 10615921122715 3:1 8 5 148 19112615 4:1 0 --6 31271911 5:1 -- -- 0 -- 6 3 14 8 6:1 ------0 -- 8 5 Source: FHWA (1992).

IV.5.3 Guard rails and impact attenuators

Obviously, it is impossible to remove all obstacles from the roadside. Those that remain can be protected by means of guard rails or impact attenuators (crash cushions). Guard rails are designed with a specific length to effectively protect a roadside object or condition considering such factors as the alignment of the road, speed of traffic and distance of the object from the side of the road. Cost-effectiveness considerations should direct decisions concerning where to install guard rails on roadsides. An example of such an approach is found in the ROADSIDE computer model (AASHTO, 1996) which is used throughout North America. A new generation model, the Interactive Highway Safety Design Model (IHSDM), is currently under development by the United States FHWA. In particular, benefit/cost analysis of road-design alternatives can be carried out in the Accident Analysis Module of this new model.

Many different types of guard rails are available that meet current crash standards. Elvik et al., (1997) summarise the findings of 20 studies on the safety effects of guard rails along the edge of the road and conclude that the installation of guard rails where none were previously placed results, on average, in 45% fewer fatal accidents, 52% less injury accidents and 18% less damage-only accidents. This data shows that guard rails are particularly effective in reducing the more severe

60 accidents. It should be noted, however, that these guard rails are developed for motorways and do not take into account specific rural road conditions. Taking this into account, the selection of the most suitable type of rails must be based on a number of factors, including traffic volume, traffic speed, traffic composition and possible impact angles. It goes without saying that just copying and applying guard rails which were developed for motorway conditions is not necessarily appropriate for rural road conditions. Of particular importance is the installation of any kind of barrier at the roadside, which in comparison to the road should not be too low (chance of roll-overs) nor too high (chance of under-run). A disadvantage of guard rails is that when they are hit, they may divert the car back into the traffic stream which, in particular on single-lane rural roads, could result in a secondary (often severe) collision. In the Netherlands, some promising research is underway to develop a safety barrier that would keep the vehicle close to the barrier in case of a collision. Initial full-scale tests with a collision speed of 50 km/h provided good results.

A concern for guard-rail designers is the simple fact that a guard rail must eventually end. This obviously presents another safety situation that must be appropriately handled. In the early years of barrier development, prime consideration was given to strength and deflection characteristics of the barrier and the protection afforded to the roadside object, with little consideration for the fact that the exposed ends of rigid or semi-rigid barriers also constituted roadside hazards. The resultant “spearing” or sudden deceleration of vehicles on exposed barrier ends led to the development of the turned-down end treatment used on W-beam, box beam and eventually the concrete safety shaped barrier. While successful to a certain extent, at least to the point of eliminating spearing and reducing the severity of collisions with blunt ends, the turned down ends were usually constructed parallel to the roadway. As a result, when a vehicle left the roadway and ran into the turned-down end, the end served as a ramp and caused the vehicle to become airborne. To avoid this type of accident, accepted practice now requires that where turned-down ends are still used, the barrier should be gradually flared away from the roadway prior to the end being turned down. In a cut situation, the end could also be buried in the roadside embankment.

Other end treatments which break away on impact, or are designed to absorb the energy of the impacting vehicle have also been developed. Recent examples of these include end treatments that either roll or cut the barrier beam as a method of dispersing the impact energy. In the case of a concrete safety shaped barrier where flaring of the turned-down end is not possible or desirable, the use of “crash cushions” or impact attenuators is recommended. These can take many forms from sand-filled barrels to various clusters of energy-absorbing materials contained within an overlapping outer casing which also provides redirectional capabilities to the attenuator for situations where it is not a head-on impact. Crash cushions can also be applied to protect solitary rigid obstacles.

IV.5.4 Break-away or flexible street furniture

Street furniture -- i.e. light poles, roadside marking posts, electricity poles and traffic and directional signs -- form a substantial part of “fixed-object” road accidents. Apart from relocating poles further away from the roadside, or protecting them by guard rails or crash cushions, break-away or flexible street furniture is an evolving technique that can contribute to the safety of rural roads (ETSC, 1998). The advantage of these materials is that they break away on impact and, because the forward motion of the vehicle is hardly affected, they tend to fall down behind the vehicle. This type of street furniture appears to have a large accident-reduction potential, in particular at speeds over 50-60 km/h (Cirillo and Council, 1986, cited by Ogden, 1996). However, this type of safety application should not be used on roads with high pedestrian and/or cyclist use because secondary

61 accidents caused by the falling pole could occur. Flexible street furniture, unlike break-away, yields progressively at impact and, as such, folds over the vehicle. Currently, break-away and flexible street furniture are mainly used for lighting poles and large directional signs. In the United States, promising activity is underway in developing break-away electricity poles. To date, two designs that meet the performance criteria identified by Ross et al., (1993) have been accepted by the FHWA. In the Netherlands, aluminium lighting poles, provided that they are not taller than 10 m, have appeared to be a good and less expensive alternative (Schoon, 1997). A deformable (patented) steel lighting pole was developed in Sweden in the 1970s. In the event of a collision, the pole does not break, but rather it bends around the vehicle.

IV.6 Signs, markings and lighting

Signs and markings fulfil the important function of warning, regulating and guiding road users. They provide extra visual information which supports the existing road design (e.g. roadside and pavement markings), complements it (e.g. local priority regulation signs) or compensates for local design elements which are below standard (e.g. curve or narrow road warning signs). In general, signs and marking need to be visible in all types of circumstances, such as rain, snow and darkness. Signs and markings also need to be legible and understandable, which means that they need to be well located, properly maintained (see also Section IV.5) and sufficiently large. The increasing number of older road users will make these two requirements even more urgent than they already are.

IV.6.1 Pavement markings

As the name suggests, pavement markings are marks on the pavement aiming to convey information to the road user in order to facilitate the driver's task. Ogden (1996) distinguishes between three types of pavement markings: longitudinal lines (centre lines, lane lines, edge lines, etc.), transverse lines (stop lines, pedestrian- or cyclist-crossing lines) and word and symbol markings (e.g. arrows for channelisation, directional information, speed limits). The latter category should be used sparingly because symbol markings are either not visible or poorly visible in adverse conditions - i.e. at night, with wet or snow-covered pavements and in heavy traffic -- and because the paint or thermoplastic material may enhance skidding, in particular of motor cycles. Central lines and edge lines are particularly helpful to position vehicles correctly in poor visibility conditions and in curves. Visibility of the lines can be enhanced by the use of retroreflective material. Using a headlight glare model (PCDETECT), a study for Transport Canada (ADI, 1992) found that edge lines are the only delineation visible to the driver under on coming headlamp glare conditions, highlighting their particular importance for safety. Several studies cited by Ogden (1996) have indicated the usefulness of longitudinal lines in reducing accidents, in particular night-time run-off the road accidents. Other studies cited by SETRA/CETUR (1992), however, indicate that on minor rural roads with low traffic volumes the introduction of edge lines and centre lines could result in higher speeds and an increase in injury accidents. Implementing pavement markings which suggest a higher level of service than allowed for by the geometric design characteristics of the road could result in a false sense of driver “comfort” and should be avoided.

Another form of pavement marking is rumble devices. Rumble devices provide an audible (grooved lines) or tactile (raised lines) cue to alert drivers and to warn them about a change in condition. They can be applied as longitudinal lines (edge line, centre line) or as transverse lines (e.g. intersection approach, speed-limit change). Rumble devices along longitudinal sections have proved to be effective in reducing run-off the road accidents that are often related to driver fatigue.

62 For example, Alberta, Canada, has started installing shoulder rumble strips on rural roads and is expecting a 20% reduction in single vehicle run-off the road accidents. The effect of transverse rumble devices is less clear, and it is suggested that, in isolation, they are insufficient to reduce speeds. The TRL in the United Kingdom (Barker, 1997) has carried out trials of various rural safety measures, including bar markings and rumble strips, as part of a wider research study into low-cost engineering measures for improving rural road safety.

IV.6.2 Roadside markings

Roadside markings include retroreflective guide posts or reflector posts located at a certain distance from the edge line with the intention to delineate the horizontal alignment. Reflector posts can be used along a road section with or without edge lines. However, they are often used on particular stretches of road only, for example to guide drivers through a curve or a bridge. Most studies show that their installation is beneficial for safety. Counter to this prevailing understanding, though, an experimental study on Finnish roads (Kallberg, 1991) found that reflector posts on roads with a speed limit of 80 km/h resulted in an increase in speed of 5 km/h during darkness and a statistically significant increase in injury accidents compared to non-treated control roads. On wider roads with higher design standards (speed limit of 100 km/h) speed choice was not affected and the overall number of injury accidents decreased, whereas the number of injury accidents in darkness increased. The latter effects, however, were not statistically significant. The Kallberg study indicates once again that care must be taken not to provide too much visual guidance on roads with relatively low design standards as it may lead to speeds which are inappropriate for the road. But, if roadside markings are restricted to particularly dangerous sites (curves, bridges, etc.), the effect on speed and accidents will be less adverse. Chevrons and specific object markers are other types of roadside markings that are used to warn drivers about particular situations such as medians, sharp curves, bridge edges, etc. They can be made of reflective material or be internally illuminated. Obviously, the material of the road-marking devices should be flexible so that it does not present a hazard to road users if hit.

Table IV.6 shows the accident reduction percentages as a result of different pavement and roadside markings. All of the figures refer to injury accidents and give the percentage change in the number of injury accidents. From this information, it is clear that none of these roadside markings significantly improve safety. The lack of safety effect is probably due to behavioural adaptation on the part of drivers, especially faster driving in the dark. It should also be noted that, depending upon the study being used as a reference, there are large differences in the reported safety effects of markings. It should be emphasized that the effects reported are largely dependent upon the specific circumstances and road characteristics in the various studies.

63 Table IV.6 Pavement and roadside markings and their effect on injury accidents

Marking Effect on safety

Ordinary edge line - 3%

Wide edge line (20 cm rather than 10 cm) + 5%

Shoulder rumble strips (broken edge line) + 2%

Reflective posts (as studied by Kallberg) + 4%

Source: Elvik et al., 1997.

IV.6.3 Road signs

Road signs provide information on local regulations, warn about upcoming dangerous situations or point out route direction or travel services. They need to be visible, legible, understandable, useful and applied consistently. All countries have nationally prescribed signs that normally take account of legibility and understandability. However, the increasing age of road users in most OECD Member countries may in particular require reconsideration of sign legibility. Visibility, usefulness and consistency of signs and signalling are causes for more concern. For instance, signs may be hidden by trees, hedges, fences or even by each other. Too many signs at a particular spot may lead to confusion or distraction rather than assisting the road user. With respect to warning signs, inconsistency can be a problem, for example, when at a particular road a warning sign is placed for some curves but not for other, similar ones. All of these are of special interest on rural roads.

IV.6.4 Street lighting

Road lighting is generally thought to reduce night-time collisions by about 25%, although individual studies show wide variation in the effect of lighting. Persaud (1992), in a review of North American and European collision experience, concluded that it appeared reasonable to expect reductions of 25-30% in total night-time collisions and 30-50% in night-time injury collisions following illumination of motorways/freeways and other principal roads. It was also stated, however, that it was unclear what circumstances -- e.g. traffic volume, road classification, vehicle speeds -- would cause the variance in these figures. Since the findings relate to motorways/freeways and principal roads, it is difficult to state whether or not they are transferable to other road types. Based on a meta-analysis of 37 public-lighting evaluation studies, Elvik (1995b) reports that the effects of road lighting vary significantly with respect to accident severity as follows: night-time fatal accidents reduced by 65%; night-time injury accidents reduced by 30%; and property-damage-only accidents reduced by 15%. Studies not specifying accident severity found a 18% reduction in night-time accidents. The meta-analysis included urban roads, rural roads and motorways. However, the results of the evaluation studies were similar for all three environments, even when controlling for accident severity.

64 Persaud (1992) also discussed the relative merit of intersection lighting as opposed to continuous roadway lighting. While the author acknowledged that the research results were rather dated, he concluded that the difference in accident rates between lighted and unlighted intersections is about 20 –– i.e. lighted intersections have accident rates nearly 20% below the accident rates at unlighted intersections.

While road lighting seems to provide reductions in the number of night-time collisions in some conditions, Section IV.5 showed that the lighting poles or masts along the roadside or in the median may also contribute to the number of injury accidents if they are not properly designed and sited.

IV.7 Maintenance and work zones

Existing roads need to be maintained in order to keep the road and its direct surroundings up to standard. Maintenance activities include those related to pavements, signs and markings as well as roadsides. Furthermore, for Nordic countries and others with severe winter conditions, winter maintenance is required to keep traffic moving safely. Whenever maintenance or reconstruction activities take place, the normal traffic situation is disturbed, generally leading to a temporary increase in accidents at and around the work zones.

IV.7.1 Pavement maintenance

Pavement-related accidents are mainly a result of low skid-resistance which leads to reduced contact between the wheels and the road surface. Skidding accidents are particularly frequent in wet weather conditions and in bends. Skid-resistance can be improved by various means such as cutting grooves into the pavement or adding a high-friction overlay. It is estimated that approximately 70% of wet-road accidents can potentially be prevented by improving skid-resistance. Another factor is the roughness of the pavement and related irregularities, such as potholes. According to Ogden (1996), in developed countries where pavements are generally well maintained, this is only a minor factor in accidents with light vehicles, but may be more important in heavy-vehicle accidents. Pavement upgrading at spots other than those with high accident incidence - - e.g. for reasons of reducing vehicle operating costs -- may have no or, initially, even negative safety effects (Hauer et al., 1994), possibly because the improvements result in higher speeds. According to the United States Transportation Research Board (TRB, 1987), routine resurfacing of rural roads initially increases dry-weather accidents by around 10%, but reduces wet-weather accidents by around 15%. They conclude that the overall effect on safety of resurfacing rural roads is small, but that it may improve the situation at roads with an extremely high number of wet-road accidents.

As mentioned in Chapter III, rutting of pavements can be serious, especially if the ruts fill with water and lead to vehicle hydroplaning. Though there are only a few references related to this subject, Lay (1981) states that hydroplaning requires a water-film thickness of 6 mm or more and speeds in excess of 80 km/h. As to rut depths to protect against hydroplaning, Haas et al., (1994) state that a rut depth of 18 mm (3/4 inch) is tolerable on collector roads. As an example of how this information is used, a recent network privatisation contract in New South Wales, Australia, required the road operator to guarantee that rut depths should not exceed 12 mm on arterial roads and 15 mm on collector roads (Haas et al., 1998).

65 IV.7.2 Maintenance of signs, markings and delineation

Signs, markings and delineation need to be maintained adequately so that they can continue to fulfil their safety function of warning, regulating and guiding road users. Signs may fade in sunshine or become illegible because of dirt or overgrown trees. Road markings and delineation may be worn off by frequent overriding of motorised traffic and their retroreflectivity may decrease. Butula (1993) distinguishes between periodic (preventive) maintenance and response (emergency) maintenance. In order to allow for adequate planning and scheduling of periodic maintenance activities and costs, he recommends that an inventory of all traffic control devices be made. Such an inventory would contain detailed information on the location, age, material, life expectancy and other pertinent data for each device. Furthermore, a schedule of routine, periodic inspections should be established to identify the following: where signs are missing; whether the devices and their supporting structures are in good condition, legible and visible; whether or not the signs are still necessary and correctly placed. It is very important that these inspections take place in both daytime and night-time conditions.

IV.7.3 Roadside maintenance

Roadside maintenance includes checking and repairing/replacing roadside safety devices and removing roadside obstructions. Preventive maintenance of roadside safety devices -- e.g. safety barriers, crash cushions, bridge barriers -- can be organised in parallel to the preventive maintenance of control devices as discussed above. In the absence of external damage caused by collisions, there is generally a long life expectancy for roadside safety equipment. To check whether there is damage that has not been reported, regular routine inspections are necessary. This, again, can be organised in parallel with the inspection of traffi-control devices. Roadside obstructions -- e.g. plants reducing visibility and sight distance, fallen trees or other objects on the shoulders -- need to be removed. In the case of plants or trees that obscure roadside signs, it may be better to relocate the sign, since the growth will continue. In practice, safety problems related to the roadside configuration or roadside furniture are generally slow to be repaired or replaced. This may partly be due to financial considerations, but it is believed that there is also a lack of awareness and technical knowledge on the part of local road authorities and maintenance supervisors about the importance of the roadside to road safety. This indicates that knowledge transfer and training in the area of roadside safety will likely contribute to better and more timely treatment of roadside hazards.

IV.7.4 Winter maintenance

In northern parts of the world, winter traffic may be disturbed by snow and ice on the road. In general, the effects of winter conditions on traffic safety are rather limited, although a local and momentary increase in accidents may occur. However, because speeds are lower, the consequences of wintertime accidents are often less severe than accidents in non-wintertime conditions. This is illustrated by Finnish accident data (Toivonen, 1998) which suggest that winter conditions mainly affect the number of damage-only accidents. The risk of all accidents on the entire Finnish road network is 36% higher in wintertime than in summertime. The corresponding figure for personal injury accidents is 4% and for fatal accidents 14%. In Finland, speed limits are reduced in winter (from 120 to 100 km/h for motorways; from 100 to 80 km/h for about 80% of other main rural roads). Furthermore, winter tyres are compulsory and 95% of the passenger cars are equipped with studded winter tyres.

66 In spite of these figures, proper maintenance contributes to the relative safety of the driving conditions in winter. An experiment in Finland (Alppivuori et al., 1995) found that reducing the normal amount of salt for de-icing roads by 80% and at the same time tripling the amount of sand, resulted in an overall increase in the number of injury accidents of around 5% compared to control roads which were treated in the usual manner. On roads with traffic volumes lower than 6 000 vehicles/day (80% of the experimental roads) the increase was around 20% (four accidents). The number of fatal accidents did not increase. On the experimental roads with reduced salting, friction levels below 0.3 were twice as common as on the control roads.

According to SETRA/CETUR (1992) it is difficult to decide on the advantages and disadvantages of different techniques to carry out winter maintenance, given the current level of knowledge on the topic. It is stressed, however, that the consistency and reliability of the winter maintenance activities on a particular road may be more important than the absolute level of maintenance.

IV.7.5 Work zones

During maintenance or reconstruction work, the normal traffic situation is disturbed. The effect on safety at and around the work zones varies widely. A study cited by Ogden (1996) of 79 roadwork projects in the United States reported an average increase in accidents of 7.5%. However, approximately one-third of the sites showed an increase of more than 50% and at 31 sites there was a decrease. The study also found that roadworks of short duration and covering a short length of road had the worst safety records.

The OECD (1989) reported that excessive speed, traffic management at the work site, adverse weather, heavy traffic and the hours of darkness are all associated with work-zone accidents. Ogden (1996) summarises that work zones should have adequate advance warning, clear and unambiguous traffic control through the work zone and that the zone be left in a safe condition when work is not in progress. Work carried out in the framework of a project funded by the European Commission (ARROWS, 1997a) resulted in an overview of currently used and innovative safety measures at work zones. However, it has been shown that the effects of different measures on behaviour and/or accidents have rarely been studied and, for the studies that have been carried out, the results are often contradictory (ARROWS, 1997b). Butula (1993) recommends weekly inspections to check the conditions and location of the local traffic-control devices which, because of the work activities, are more prone to disruption and dirt. With respect to speed, the OECD (1989) states that the speed limits in work zones should be realistic, be supported by appropriate accompanying measures and not rely solely on signalling. If road works are of such a magnitude that they require a diversion of traffic, the alternative route is often of a lower standard and not equipped to deal with high flows of through traffic. In this situation, diverted drivers may not sufficiently adapt their behaviour to the new (lower standard) road conditions. Careful attention therefore has to be paid to the safety of the diversion route. For example, clear and unambiguous directional signalling, warning signs, posted speed limits and other, temporary measures may all help to maintain an acceptable level of safety.

IV.8 Vulnerable road users: pedestrians and bicyclists

The number of bicyclists, mopeds and pedestrians on rural roads varies widely across countries. The level of this kind of traffic is determined mainly by the physical -- e.g. hilliness,

67 available facilities -- and cultural -- e.g. the status of the bicycle as a means of transport, walking or jogging for physical fitness -- characteristics of the country. Generally, rural roads have very low flows of pedestrians, except perhaps in areas with some commercial and recreational activities. Bicyclists and light mopeds can be much more common on rural roads. The consequences of accidents between motorised vehicles and vulnerable road users are generally very serious because of the large differences in speed and mass. For example, when the vehicle is travelling at a speed of 65 km/h, 85% of the pedestrians involved in a collision with a vehicle die. In comparison, only 5% of pedestrians die if the vehicle is travelling at a speed of 30 km/h (ETSC, 1995). Whereas the majority of bicycle accidents happen in urban areas, the fatality rate is much higher outside the urban areas. The two major bicycle accident categories are those related to crossing the road and those related to the mixed use of the same road section. TRL report 310 on rural cycling (Gardner, 1998) highlights the increased risk to cyclists of fatal injuries if involved in an accident on rural roads, and the wide-spread locations of cyclist accidents which makes single-site treatment more difficult. The higher speeds of traffic on rural roads is also a factor.

IV.8.1 Road sections

By keeping pedestrians and bicyclists spatially separated from fast-moving motorised traffic, the majority of serious accidents resulting from mixed use of the same road can be prevented. From a safety point of view, the best solution is separate lanes for bicycle and pedestrian traffic. Special attention has to be paid to situations where the bicycle/pedestrian lanes cross the main road (see below) and where the separate (bicycle) lane merges with the main road, for example when entering a village. Bicycle and pedestrian lanes can be provided on one side or on both sides of the main carriageway. ERSF (1996) advises that one-way bicycle lanes should be provided on both sides of the main road in the following cases: • if there is a large amount of bicycle traffic; • if the road passes through several villages in close succession; or • in an area with scattered settlements.

Denmark and the Netherlands discourage the use of bi-directional bicycle lanes because car drivers do not expect cyclists approaching from the “wrong” direction and because it requires extra crossing manoeuvres for the bicyclists.

IV.8.2 Road crossings

The most effective way to reduce crossing accidents is to provide grade-separated crossing facilities. This is, however, a very expensive solution that is only cost effective at locations with a very high frequency of accidents. Furthermore, the use of underpasses and overpasses is not very popular with pedestrians and bicyclists, especially if their use requires more time and physical effort than an at-grade crossing. Constructing an underpass or overpass for the motorised vehicles to overcome this problem is even more expensive, in particular for existing roads. Traffic signals can be an alternative for crossing major roads, although this may have adverse safety effects for the motorised traffic on the major roads.

ERSF (1996) lists three safety principles related to non-signalised at-grade crossing facilities for pedestrians and bicyclists:

68 • to facilitate co-ordination of movement between drivers of motorised vehicles and vulnerable road users by enhancing mutual visibility; • to shorten the crossing exposure for vulnerable road users by minimising conflict lengths; and • to reduce the speeds of vehicles.

Minimising conflict lengths can be accomplished by road narrowings -- this also has a speed-reducing effect -- or by a median refuge. Whether these are feasible solutions depends on, among many other things, the balance between the motorised and non-motorised traffic flow. With respect to bicycle lanes parallel to the main road and crossing a minor road, ERSF (1996) distinguishes two possible design lay-outs. One approach is to continue the bicycle path or lane parallel to the major road with bicycles having priority over traffic from the minor road. This design is only an alternative if the bicycle lane is a one-way lane only, but even then this design may give rise to conflicts between bicyclists and motorised traffic turning from the major road into the minor road. This type of conflict can be reduced by tracing the cycle lane directly adjacent to the carriageway from about 30 m before the intersection. The second approach provides a bicycle lane that is slightly staggered away from the major road -- approximately the length of one passenger car -- and bicyclists have to yield for traffic on the minor road, including traffic that has just turned from the major road into the minor road. According to the ERSF report, the second approach must always be applied in the case of a two-way bicycle lane, but may also be preferred on a one-way bicycle lane. Roundabouts are also an option, though the positive safety effect of a roundabout for bicyclists is smaller than for other road users. However, a roundabout is still substantially safer than the conventional four-armed intersection (see Section IV.4.5). Research and discussion are still going on with respect to the best place for bicyclists on a roundabout and, in the case of separate traffic lanes, about the safest priority regulation.

IV.9 Identifying (potential) safety problems and solutions

The majority of existing rural roads were built when safety was not an issue. Whereas safety aspects have become increasingly important during the last decades, it still too often only an implicit argument that has to compete with arguments for efficiency, comfort and overall cost. As a consequence, existing roads and, often, new road schemes lack the basic elements which are known to improve the safety of their users. A number of often complementary methods are available which aim to check the safety of existing roads, of new road schemes and of major reconstruction projects in a systematic and explicit way. Experience shows that identified problems can often, though not exclusively, be resolved by low-cost, high-benefit measures.

IV.9.1 Road safety inspections

Road safety inspections provide for a systematic and comprehensive overview of the safety characteristics of a road that takes account of its current function and use. The results, a “catalogue” of general safety deficiencies on existing roads, may form the basis of safety improvement programmes. Road safety inspections must be carried out by road safety experts. They generally cover a large part of the road network and are initiated with or without a bad accident record. Inspections of single high-risk locations or specific high-risk stretches of road are generally known as black spot analysis (see below). Road safety inspections can be carried out on the basis of common

69 sense and expert judgement or by comparing the existing situation to the latest guidelines (Slop and Catshoek, 1995), although in practice it is often a combination of the two. The level of qualification and expertise of the inspectors is clearly of the utmost importance for the success of this approach. The overall effect of road safety inspections is difficult to assess and is largely dependent on the measures which are taken after a particular safety problem has been identified.

IV.9.2 Black spot and black route analysis

Black spot analyses are carried out if the accidents in the road network appear to cluster on particular spots in the network, the so-called “black spots” (PIARC, 1995). The analysis and subsequent identification of appropriate treatment require well-developed local accident data systems. Box IV.3 illustrates the effectiveness of this type of analysis.

Box IV.3 Black spot/route programmes are effective

In Australia, the federal government, in co-operation with the state governments, targets locations and road lengths with a serious crash history. The 1996-2000 Black Spot Road Safety Funding Programme represents the second programme of its kind that addresses such sites on the road network, with funds being maintained in real terms over the life of the programme. The initial programme was evaluated by the BTCE (1995) which concluded that, from a random selection of sites funded under the programme, the average benefit-cost ratio exceeded 4.0. In the current Programme, there is a 50:50 split of funds between rural and urban locations. This allocation of funds reflects the distribution of road fatalities in Australia. Further, up to 20% of funds may be directed to preventive road safety by addressing potentially serious crash sites based on a road safety audit.

ETSC (1996) summarised that an effective local accident data system should allow for the following actions:

• the retrieval of accident data for each particular site; • the identification of sites with a high number of accidents; • the identification of sites with common accident types; and • the ability to relate the above information to police road accident report forms to enable details of particular accidents to be followed up.

Such local accident data systems allow a detailed assessment not only of the number of accidents, but also of the accident types at a certain location. To proceed further, the high risk locations need to be prioritised and the common factors in accidents and accident occurrence at the selected locations identified. To get good insight into the situational factors that may explain the high accident numbers, a site visit is an essential element of the analysis procedure. In a rural environment the black spot approach is mainly applicable on high-volume roads because accidents on low-volume roads tend to be too scattered to allow for the identification of accident clusters on a particular spot. However, a similar approach can be applied to stretches of road rather than to specific locations. For example, in the United Kingdom, accident rates of a particular road are compared with the average rate of similar road types and if the rate exceeds a predetermined intervention level, further action is taken. As for road safety inspections, the effectiveness of black spot and “black route” analyses depends mainly on the effectiveness of the subsequent remedial measures. A thorough analysis of the problem, however, is a prerequisite for the identification of effective treatment.

70 IV.9.3 Safety audits

A safety audit is a formalised and standardised procedure to independently judge the potential safety effects of road schemes. A safety audit differs from a safety inspection in that an audit is much more highly focussed on addressing potential safety problems during the design process of a defined road project. The primary output of a safety audit is recommendations to resolve identified likely problems. A safety audit therefore aims at crash prevention rather than responding to those crashes that have already happened. Originating in the late 1980s in the United Kingdom, an increasing number of countries -- Australia, New Zealand and a number of Northern European countries -- started to apply safety audits. The important elements of a safety audit are that it is a formalised procedure, that the process is applied independent of the designer or road authority and that safety is systematically designed into each project. Both AUSTROADS (1994) of Australia and the Insitution of Highways and Transportation (1997) of the United Kingdom have documented safety audit procedures.

A safety audit procedure usually starts in the early stages of the project planning, e.g. the feasibility study phase or the draft design phase. An independent auditor or audit team looks at the design plans in detail, generally using detailed check-lists, with safety as the sole guiding principle. They then report on the potential safety problems which may arise from the plans and formulate recommendations for solving the problems. Subsequent safety audits are carried out at the detailed design phase and prior to the opening of the facility. In some cases, the audit is repeated once more when the road is in-service. It is also possible to apply the safety audit procedures and check-lists for general road safety inspections as described above. By applying safety audits, remedial works that are often more expensive after construction are largely prevented. As well, the awareness of the safety consequences of planning and design decisions may improve among all those who are involved.

Whereas it is difficult to estimate the safety benefits from safety audits in a reliable way, the experiences thus far have been very positive and, given the relatively low cost -- estimated to be between 1 and 4% of the total project cost -- safety audits seem to be cost effective in the majority of cases. For example, in Denmark, a benefit-cost analysis on 13 different schemes comparing all costs involved in the application of an audit (extra time investment, changes in construction costs) with the savings due to expected accident reduction, showed an average first year rate of return of 146%, making the audit a highly profitable instrument (Danish Road Directorate, 1997).

IV.9.4 Low-cost road and traffic engineering measures

“Low-cost road and traffic engineering measures comprise those physical measures, taken specifically to enhance the safety of the road system, that have low capital cost, can be implemented quickly and offer high ratios of benefit to cost. Achieved ratios of benefit to cost from three upwards to double figures are widely reported, and many schemes pay for themselves in casualty savings within a year” (ETSC, 1996).

Low-cost measures (LCM) can be applied at high-risk sites or along high-risk road sections where accidents may not be concentrated at particular sites. A prerequisite for the success of the LCM approach is the availability of an effective local accident data system which allows for the identification and analysis of “black spots” and “black routes”. By studying the accident data, identifying common accident factors and relating the specific site characteristics to the accident

71 patterns, it must be decided whether LCM offers a viable solution and, if so, which types of measures are most likely to resolve the problem. A further requirement is the systematic evaluation and dissemination of the results, which allows for a continuous update of the information on effectiveness of specific measures in specific circumstances. Obviously, to do so, it is important to report not only the success stories, but also the less successful activities.

In the previous sections, a considerable number of low-cost measures were mentioned, such as the installation of guard rails, the removal of obstacles from the roadside, improving the visibility of markings and signs, speed-reducing measures at village entrances, upgrading advance warning for narrow curves, and many others. Obviously, the determination of whether or not a measure is “low-cost” or cost effective depends on the available budget, local/national price characteristics and whether or not improvement schemes can be combined with regular road maintenance activities. Although it must be clear that a more structural, network-wide or area-wide -- and hence more expensive -- approach is required to come to an inherent, sustainably safe road environment, low-cost measures can contribute substantially to the safety of those parts of the network where it is needed most.

IV.10 Conclusion

It is rarely the case that an accident can be explained by only one cause. Generally, an accident is the result of a combination of human factors, vehicle factors and road-design factors. Thus, safety work must focus jointly on all three elements. Road design, however, offers a very good opportunity to improve road safety, since safe road design can reduce:

• the incidence of human error; • the incidence of human error or vehicle error resulting in a collision; and • the severity of the consequences of a collision, if it cannot be avoided.

Knowledge about safe road design is developing rapidly, although it is still incomplete. Whereas for some design elements reliable quantitative estimates of the safety effect are available, for many other elements only the direction of the effect is known. Furthermore, the ultimate safe road design requires the application of the optimal combination of all different design elements. There is still a lot of work to do in this area. A systematic research approach with generally accepted research methodologies and tools would markedly increase the speed of knowledge development as well as the reliability and usefulness of the results. In addition, specific attention to safety as a basic design element in university-level road engineering courses is recommended.

While safety is increasingly understood to be an important criterion in road design, it is still too often of secondary importance in daily practical road and traffic engineering work. Explicit attention to safety means that safety must be considered in all stages of road and traffic engineering, starting at the level of the decision to build or rebuild a road, followed by the overall planning stage, the design stage, construction, during operation and maintenance. The earlier in the process safety is considered, the smaller the need for sometimes expensive remedial work afterwards. An increasing number of countries apply road safety audits -- i.e. a systematic and independent check of the potential safety consequences during planning, designing and (re)building road infrastructure -- that aim to prevent accidents rather than responding to those that happen. The experiences are very positive and the safety effects, though difficult to quantify, are generally reported to be positive. Wide use of safety audits at the local, regional and state levels is recommended.

72 The basis of a safe road design is a consistent, hierarchical road network, in which each road category has a particular function to fulfil (see Chapter VIII). The design characteristics of a road need to be in accordance with its function, and as such automatically elicit the appropriate expectations and behaviour of road users. In other words, the road environment needs to act as “positive guidance” for road users. In some cases, this may mean that a road needs to be upgraded (including the construction of parallel facilities for non-motorised traffic and access purposes), for example to fulfil its role for motorised long-distance traffic. In other cases, a road may need to be downgraded, for example to reach a speed level of motorised traffic which is appropriate when sharing the road with non-motorised traffic.

A very important and often overlooked aspect of the final safety level of the road is the lay-out of the direct surroundings of the road -- i.e. the roadside. A sufficiently large clear zone and appropriate protection of obstacles which cannot be removed have tremendous potential to reduce the number of accidents and reduce the severity of those that cannot be avoided. This is particularly true for rural roads where a large proportion of the accidents are single vehicle run off the road accidents. More than is currently the case, the roadside must be considered as part of the road and as such be seen as the explicit responsibility of road engineers.

Because the road network outside built-up areas is so extensive, the issue of cost-effectiveness naturally arises as to “where” and “when” measures should be applied to improve safety. It is therefore of primary importance to determine what the function of the roads are because this will define the context for further measures. There are then two approaches. On the lower order road network with low-volume roads where there is only local traffic reaching its destination, the work method could be, if road maintenance is necessary, to ensure a defined minimum safety level before any maintenance is undertaken. In the “maintenance tempo”, safety improvements could then be carried out, preferably incorporating an approach that addresses a comprehensive network or related stretches of roads in order to prevent discontinuity.

On the higher order road network where there is a question of a high share of through traffic, high volumes and relatively high speeds, an analysis of the risk -- i.e. accidents that have happened -- is desirable on the basis of which spot-related and route-related measures could be designed and implemented. In this instance as well, infrastructural improvements should explicitly deal with access control, meet design consistency rules and not cause discontinuities in the course of the roads. Furthermore, the question should always be raised as to whether infrastructural improvements of the road will lead to higher driving speeds and, therefore, not lead to the expected safety gains.

73 BIBLIOGRAPHY

AASHTO (1996), Roadside Design Guide, American Association of State Highway and Transportation Officials, Washington, D.C.

ADI (1992), Driver Visibility Requirements and Reflective Standards, prepared for Transport Canada, Ottawa, Ontario.

ADI (1989), Passing Manoeuvres and Passing Lanes: Design, Operational and Safety Evaluations, prepared for Transport Canada, Ottawa, Ontario.

ALPPIVUORI, K., LEPPÄNEN, A., ANILA, M., and MÄLELÄ, K. (1995), Road Traffic in Winter; Summary of Publications in the Research Programme, Finnra Reports No. 57/1995, Helsinki: Finnish National Road Administration, Helsinki.

ARROWS (1997a), Road Work Zone: Typology, Standards and Practices; Parts A and B, Deliverable 1, EC Research Project RO-96-SC.401.

ARROWS (1997b), Road Work Zone: Behavioural Studies, Accident Studies and Research Methods, Deliverable 2, EC Research Project RO-96-SC.401.

ASIAN DEVELOPMENT BANK (1996), “Road Safety Guidelines for Asia and Pacific Region”, Draft 2, TRL, Berkshire.

AUSTROADS (1994), Road Safety Audit, AUSTROADS, Sydney.

BARKER, J. and HELLIAR-SYMONS, R. (1997), “Count-down Signs and Roundel Markings Trials”, TRL Report 201, Transport Research Laboratory, Crowthorne, Berkshire.

BARKER, J. (1997), “Trials of Rural Road Safety Engineering Measures”, TRL Report 202, Transport Research Laboratory, Crowthorne, Berkshire.

BARKER, J., FARMER, S. and NICHOLLS, D. (1998), “Injury Accidents on Rural Single-Carriageway Roads, 1994-95: An Analysis of STATS 19 (UK) Data”, TRL Report 304, Transport Research Laboratory, Crowthorne, Berkshire.

BRANNOLTE, U., DILLING, J., DURTH, W., HARTKOPF, G., MEEWES, V., ROHLOFF, M., SCHLIESING, H., and STIEVERMANN, P. (1992), Einsatz von Zwischenquerschnitten, Bundesanstalt für Strassenwesen, Bergisch Gladbach.

BTCE (1995), Evaluation of the Black Spot Program, Report 90, Australian Government Printing Service, Canberra.

74 BUTULA, R.J. (1993), “Infrastructure Maintenance - Traffic Control Devices”, ITE: The Traffic Safety Toolbox, Institute of Transportation Engineers, pp. 171-178, Washington.

CIRILLO, J.A. (1993), “Roadside Safety”, ITE: The Traffic Safety Toolbox, Institute of Transportation Engineers, pp. 149-155, Washington, D.C.

DANISH ROAD DIRECTORATE (1994), Road Safety at 82 Danish Roundabouts, Danish Road Directorate, Road Safety and Environment Department, Copenhagen.

DANISH ROAD DIRECTORATE (1997), Road Safety Audit - Evaluation of the Pilot Project, RSA- information 2/97, Danish Road Directorate, Copenhagen.

ELVIK (1995a), “The Safety Value of Guard-rails and Crash Cushions: A Meta-analysis of Evidence from Evaluation Studies”, in: Proceedings of the Conference Road Safety in Europe and Strategic Highway Research Program (SHRP), No. 2a, Part 2, p. 190-230, VTI Swedish Road and Transport Research Institute, Linköping.

ELVIK (1995b), A Meta-analysis of Evaluations of Public Lighting as an Accident Countermeasure, paper presented at the 74th Annual Meeting of the Transportation Research Board, Transportation Research Board, Washington D.C.

ELVIK, R., BORGER, A., and VAA, T. (1997), Traffikksikkerhetshåndbok, Institute for Transport Economics TØI, Oslo.

ERSF (1996), INTERSAFE, Technical Guide on Road Safety for Interurban Roads, European Road Safety Federation, Brussels.

ETSC (1995), Reducing Traffic Injuries from Excess and Inappropriate Speed, European Transport Safety Council, Brussels.

ETSC (1996), Low-cost Road and Traffic Engineering Measures for Casualty Reduction, European Transport Safety Council, Brussels.

ETSC (1998), Forgiving Roadsides, European Transport Safety Council, Brussels.

FHWA (1981), A User’s Guide to Positive Guidance (2nd edition), Report No. FHWA-TO-81-1, Federal Highway Administration, Federal Highway Administration, Washington, D.C.

FHWA (1992), Safety Effectiveness of Highway Design Features (Vol. I-VI), Publication No. FHWA- RD-91-044 to 049, Federal Highway Administration, Federal Highway Administration, Washington, D.C.

FHWA (1993), The 1993 Annual Report on Highway Safety Improvement Programs, Federal Highway Administration, Washington, D.C.

GARDNER, G. and GRAY, S. (1998), “A Preliminary Review of Rural Cycling”, TRL Report 310, Transport Research Laboratory, Crowthorne, Berkshire.

HAAS, R., HUDSON, W.R, ZANIEWSKI, J. (1994), Modern Pavement Management, Krieger Publishing, Malabar, Florida.

75 HAAS, R., MEYER, F., COWE-FALLS, L. (1998), “Assessment Framework of Infrastructure Maintenance”, Proceedings of the Annual Conference of the Transportation Association of Canada, Regina, Saskatchewan.

HAUER, E. and PERSAUD, B. (1996), Safety Analysis of Roadway Geometry and Ancillary Features, Transportation Association of Canada (TAC).

HAUER, E., TERRY, D. and GRIFFITH, M.S. (1994), “The Effect of Resurfacing on the Safety of Two-lane Rural Roads in New York State”, paper presented at the 73rd Annual Meeting of the Transportation Research Board, Paper No. 940541, Transportation Research Board, Washington D.C.

INSTITUTION OF HIGHWAYS AND TRANSPORTATION (1997), Guidelines for the Safety Audit of Highways, Institution of Highways and Transportation, London.

LAY, M.G. (1981), Source Book for Australian Roads, Australian Road Research Board, Vermont South, Victoria.

MAYCOCK, G., BROCKLEBANK, P. and HALL, R. D. (1998), “Road Layout Design Standards and Driver Behaviour”, TRL Report 332, Transport Research Laboratory, Crowthorne, Berkshire.

MCGEE, H.W., HUGHES, W.E., DAILY, K., and BELLOMO-MCGEE INC. (1995), Effect of Highway Standards on Safety, NCHRP Report No. 374, Transportation Research Board, Washington, D.C.

NCHRP (1997), “Strategies for Improving Roadside Safety”, Research Results Digest 220, Transportation Research Board, Washington D.C..

NILSSON, G. and WENÄLL, J. (1997), “Päkörning av belysningsstolpar och andra härda föremal i vägmiljön, (Collisions with Lamp Posts and Other Hard Obstacles in the Road Environment)”, VTI Meddelande No. 825, Swedish National Road and Transport Research Institute, Linköping.

OECD ROAD TRANSPORT RESEARCH PROGRAMME (1989), Traffic Management and Safety at Highway Work Zones, Paris.

OGDEN, K.W. (1996), Safer Roads, a Guide to Road Safety Engineering, Avebury Technical, Aldershot.

PAJUNEN, K. and KULMALA, R. (1988), “Traffic Safety on Overtaking Lanes in Finland”, (Ohituskaistojen turvallisuus), VTT Research Report 678, Espoo, Finland.

PEO (1997), Highway 407 Safety Review, Professional Engineers of Ontario, Toronto.

PERSAUD, B. (1992), Roadway Safety - A Review of the Ontario Experience and Relevant Work Elsewhere, Ministry of Transportation, Research and Development Branch, Downsview, Ontario.

76 RONA (1987), Richtlijnen voor het ontwerpen van niet-autosnelwegen; “Hoofdstuk II: dwarsprofielen.” (Guidelines for the Design of Non-motorways; Chapter II: cross section), Sdu Uitgeverij, The Hague.

ROSS, H.E., SICKING, D.L., ZIMMER, R.A., and MICHIE, J.D. (1993), Recommended Procedures for the Safety Performance Evaluation of Highway Features, NCHRP report 350, National Research Council NRC, Transportation Research Board TRB/National Academy Press, Washington D.C.

RUMAR, K. (1985), “The Role of Perceptual and Cognitive Filters in Observed Behavior”, in: L. Evans and R.C. Schwing (eds) Human Behaviour and Traffic Safety, Plenum Press, pp.151-165, New York.

RUYTERS, H.G.J.C.M., SLOP, M. and WEGMAN, F.C.M. (1994), Safety Effects of Road Design Standards, Report R-94-7, SWOV Institute for Road Safety Research, Leidschendam.

SANDERSON, R.W. (1996), Canadian Road Safety and Public Highway Infrastructure, Report No. TP 12801E, Transport Canada, Ottawa.

SAFESTAR (forthcoming), Safety Evaluation of Different Kinds of Cross-sections on Rural Two-lane Roads, Deliverable 4.1, EC Research project RO-96-SC.203.

SCHOON, C.C. (1993), Ongevallen op rotondes II, (Accidents at Round Abouts, Part II) SWOV Road Safety Research Institute, Leidschendam.

SCHOON, C.C. (1994), Road Design Standards of Medians, Shoulders and Verges, Report A-94-9, SWOV Institute for Road Safety Research, Leidschendam.

SCHOON, C.C. (1997), Roadside Design in the Netherlands for Enhancing Safety, Report D-97-20, Leidschendam: SWOV Institute for Road Safety Research, The Netherlands.

SETRA (1994), Aménagement des routes principales, Service d’Etudes Techniques de Routes et Autoroutes, Bagneux.

SETRA/CETUR (1992), Sécurité des routes et des rues, le Services d’Etudes Techniques de Routes et Autoroutes, Centre d’Etudes des Transports Urbains, Bagneux.

SLOP, M. and CATSHOEK, J.W.D (1995), Recommended Safety Measures for Application on Interurban Roads in the Short Term, Report R-95-18, SWOV Institute for Road Safety Research, Leidschendam.

SWOV (1994), Infrastructure: Design and Road Safety, Part 1: Summaries and Statement, report on an OECD Workshop held 15-18 November in Prague, Report D-94-14 I, SWOV Institute for Road Safety Research, Leidschendam.

TOIVONEN, S. (1998), Safety Problems of Rural Roads, Finnish National Road Administration (Finnra), Helsinki.

TRB/NRC (1987), Designing Safer Roads; Practices for Resurfacing, Restoration and Rehabilitation, Transportation Research Board, National Research Council, Washington D.C.

77 TRL (1994), “Speed Reduction in 24 Villages: Details from the VISP Study”, TRL Project Report 85, Transport Research Laboratory, Crowthorne, Berkshire.

TRL (1996), “Traffic Calming on Major Roads: The A49 Trunk Road at Craven Arms, Shropshire (UK)”, TRL Report 212, Transport Research Laboratory, Crowthorne, Berkshire.

ZEGEER, C., REINFURT, D., NEUMAN, T., STEWART, R. and COUNCIL, F. (1990), Safety Improvements on Horizontal Curves for Two-lane Rural Roads: Informational Guide, Highway Safety Research Center, University of North Carolina, Chapel Hill, N.C.

78 Chapter V

ENFORCEMENT

V.1 Introduction

Enforcement is one of several factors that contribute to the efficiency and safety of the traffic system. For instance, the traffic system has often been described by the three Es: Engineering, Education and Enforcement. Undoubtedly, other factors can be added to the system such as the socio-economic and cultural environments of the road user, the natural environment -- i.e. climate, urbanisation, terrain, etc. -- and the legal system (see Figure V.1). However, in terms of rural road safety, enforcement can have a significant impact on both the current level of safety as well as creating positive trends for improved levels of road safety in the future by influencing road-user behaviour. This chapter therefore describes the importance of enforcement, the enforcement mechanisms and a variety of factors that influence enforcement activities. A detailed discussion on enforcement practices is also presented.

Figure V.1 System components and interactions

Social and cultural environment Road users

Laws and regulations Road user ROAD behaviour SAFETY Police enforcement Road network Vehicles Built environment

V.2 Effectiveness and limitations of enforcement

Traffic violations are an exception in our legal and enforcement system. With most crimes and misdemeanours, society is dealing with a small minority of violators whom the police set out to apprehend. There is wide public support for this approach when dealing with murder, theft, drugs, physical violence, and other crimes. Police activity and success are generally limited only by the resources that can be applied and by the established priorities. When dealing with traffic violators, on the other hand, the public at large is involved. It is not simply the case of removing a small group of

79 drivers from the traffic stream in order to solve the problem. Nearly all road users violate the regulations to one extent or another by sometimes exceeding the speed limit, driving in the wrong lane, not slowing down at an intersection or other infractions.

A comprehensive study on the deterrence of high-speed driving from a criminological perspective (Corbett et al., 1998) considered the internal and external measures to prevent speeding. These included the usual means of enforcement, engineering measures and penalties plus a wider look at more innovative solutions including smart cards. The main conclusion was that the problem of speeding is a socio-cultural one and cannot be located solely with the individual driver; in the Westernised car culture with everything geared towards speed and mobility, the public sector must take the lead in attempting to modify drivers’ attitudes as to the desirability of high speed. Further, the strong impression was that those who drive fastest do so partly because they feel most invincible and most in control. Research was encouraged to understand how this misconception of invulnerability arises and how it can be changed.

The traffic enforcement system is set within a social context that requires public support through the media and the political system. The technology currently exists where electronic licence plates, when attached to a vehicle and supported by appropriate roadside equipment, could allow for the continuous registration of offences such as speeding, not stopping at a traffic signal, crossing a white line, and illegal parking. Widespread application of such a system could be highly cost effective and would considerably reduce the role of manned police. The fact that they have not been implemented, even on an experimental basis, points to the fact that it does not now enjoy public support.

Along similar lines, but less drastic, the police could, without too much trouble, enforce the speed limits by ticketing every violator. Within a short time, millions of tickets would be issued, the courts would be flooded and public reaction would be such that the process would be halted. It follows that police enforcement of traffic regulations has to be conducted carefully on the basis of wide public support.

The enforcement process and its aims are described in Figure V.2. Police presence in general and the various means and methods of active enforcement create a certain objective risk of apprehension. This varies from offence to offence, depending on the amount of enforcement applied and on the methods used. The objective risk of apprehension, together with publicity, creates within the road user a subjective probability of apprehension. The subjective probabilities of apprehension, together with the expected punishment for violators which is meted out by the judicial process, constitute the deterrence effect.

The rational road user acts on the principle of utility. As long as the utility to be derived from a certain behaviour is greater than the expected punishment, he will not change his or her behaviour (Bjornskau and Elvik, 1992). The utility is made up of time savings, pleasure -- e.g. in fast driving -- or “inertia” reflected in a certain resistance to changing a course of action, as for example, the need to brake at a pedestrian crossing or traffic signal. The deterrence is made up of the subjective risk of apprehension multiplied by the expected punishment and, in some cases, a certain probability of causing an accident, although this is almost never explicitly considered.

The process described in Figure V.2 creates the deterrence that, together with driver education and training, can change road-user behaviour. In time, and when applied on a large enough scale and consistency, behaviour changes can lead to shifts in the norms of behaviour. At this stage,

80 the changed behaviour is internalised and will be applied even without deterrence measures. It is thus envisaged that the enforcement at some stage becomes superfluous or can be reduced because drivers have internalised their new behaviour and act accordingly. Such changes can be observed in the use of seat belts and motorcycle helmets as well as, to some extent, driving under the influence of alcohol.

Figure V.2 The enforcement process and its components

Active Police presence Publicity enforcement

Subjective probability Punishment of apprehension

Deterrence Education

Change behaviour

Change norms

Change in accidents/safety

In the final stage, as depicted in the diagram, there is a link between changed norms of behaviour and a reduction in accidents.

V.2.1 Associated factors -- laws, regulations and punishment

As described in Chapter III, enforcement is one element that interacts with a wide variety of other engineering, socio-economic and environmental factors to generate certain types of road-user behaviour. The norms of behaviour, or normal behaviour, are set by society through its administrative and legal procedures. As well, in the traffic environment, traffic laws and regulations form the background against which behaviour is judged. When laws are such that they are accepted by the wide majority of society, they become, to a large extent, self-enforcing. In other words, when laws are the norm, road users accept them and little police effort is required to enforce them.

In the case of speed limit regulations, on the other hand, a large proportion of the public -- sometimes the majority -- does not behave according to the law. In such a case, it could be asked if the law should not be adapted to be in better accordance with “normal” behaviour. If, however, for reasons of safety, environmental considerations or others, it is decided that the law should not be adapted to match behaviour, the system falls back on enforcement and public awareness to induce correct behaviour. It is in such situations that the limitations of enforcement become most apparent. In these conditions and in general, enforcement can only be effective if it operates in a supportive

81 environment of laws, regulations and a sensitive penal system. These combined factors act to create the deterrence effect of police enforcement, both on the individual level and on society at large.

Strong enforcement cannot be effective if it operates in combination with loose laws and small fines, or fines that only have to be paid after many years. The desired degree of deterrence is, therefore, achieved by a mix of the above-mentioned ingredients. The generally held view on the effectiveness of enforcement is that a combination of factors should be considered. The deterrence effect is the combined result of enforcement and punishment as manifested in the subjective risk of apprehension. In risk of apprehension, it is certainty that produces the greatest effect. Certainty is a question of enforcement intensity, and will be further explored in the following section on automatic enforcement. Swiftness of punishment seems to be the most important factor for effective enforcement (Shinar, 1978; Seipel, 1992).

Severity of existing punishment does not seem to have a great proven effect. However, stricter penalties will reduce the level of enforcement required. (Mäkinen et al., (1992); Seipel (1992); Bjornskau and Elvik (1992)) More work is needed to examine the effect of punishment severity -- i.e. financial impacts, loss of licence, etc. -- as a deterrent.

Many countries have installed penalty-point systems whereby drivers accumulate points in association with offences committed. However, little is known about the proven effectiveness of point systems mainly because it is difficult, if not impossible, to measure. It appears likely that the knowledge of the possible sanctions to be activated when a certain number of points have accumulated can also serve as a deterrent.

Box V.1 Enforcement and publicity

It has been well established that the effectiveness of enforcement can be enhanced by giving it the right amount of publicity. In the past this was applied more to combined enforcement-publicity campaigns that were limited in time and focused on such things as seat-belts, drink-driving or fatigue (on holidays). These joint efforts still enjoy wide support both from the professional and the general public. In the past few years, a new practice is developing especially in relation to speeding. More and more countries are using the joint tactic of increased enforcement together with prior notification in which the public is forewarned about speed measurements on certain stretches of road. This practice is widely used in the Nordic countries, the Netherlands, the United Kingdom, Australia, Israel and others.

This chapter deals mostly with the enforcement duties of the police. It should, however, be realised that police duties extend far beyond the enforcement of traffic laws and regulations. The police have a prime duty with respect to alleviating traffic congestion and maintaining the orderly flow of traffic, they are involved in all aspects of dealing with traffic crashes, generally are in direct contact with the traffic engineering community and have a major role to play in educating the general public concerning road safety through community programmes.

V.3 The enforcement mechanism

This section delves into the various methods of enforcement by discussing the following issues in the context of available knowledge: Which behaviour should be targeted for enforcement? How to enforce? How much to enforce? Where and when to enforce (i.e. operational plans)?

82 V.3.1 What to enforce?

Most of the enforcement literature deals with a limited number of behaviour types including exceeding speed limits, driving while intoxicated, the use of seat-belts, the use of motorcycle helmets and integrated programmes applied to cities or areas. Other issues, including driver and vehicle licence checks, parking offences and driver behaviour at intersections, are mentioned, but not to any wide extent. One of the most difficult issues to enforce, but one which is a high contributor to fatalities in some countries, is driver fatigue.

The decisions on the behaviour to be enforced depend, to a large extent, on the view of the authorities in general, and the police in particular, as to the contribution of those kinds of behaviour to accidents. The general police practice is to associate a violation with each accident recorded. These violations -- such as excessive speed, lack of attention, not yielding, following too closely -- are, in some cases, also termed (on official forms) accident causes. It is accepted that police must form their opinion as to the violations committed in order to prosecute drivers involved in or responsible for road crashes. From this viewpoint, the type of enforcement that will be termed “correlative enforcement” evolves (Anon, 1985). According to this concept, police will enforce certain behaviours in relation (or in proportion) to their share in the accident statistics. If, for example, 20-25% of the accidents are associated with “not giving way at an intersection,” -- i.e. not stopping at a stop sign or failing to yield for main road traffic -- then it follows that a similar effort should be put into enforcing this type of behaviour.

In a different manner, but thinking along similar lines, if a certain number of accidents occur on certain roads, and if by certain standards, this number is too high, such roads will be termed hazardous or high-risk locations. One view held by the police is that enforcement should be focused on such locations. This raises another associated issue of whether enforcement should be “repressive” by catching as many offenders as possible and punishing them, or “preventive” by interacting -- i.e. through visible patrols, stopping drivers, issuing warnings or perhaps rewarding non-offenders -- with as many road users as possible. In practice, efficient enforcement will almost always be a mix of both types.

V.3.2 How to enforce (enforcement tactics)?

Much research has dealt with the relative effectiveness of various enforcement tactics. Many of the experiments dealt with speed enforcement, but others also treated the use of seat-belts and driving while intoxicated. Good reviews of enforcement tactics can be found in Shinar and McKnight (1984), Goldenbeld (1993), Riedel et al., (1988), Makinen and Syvanen (1990), Armour (1984) and Zaal (1994).

In speed-limit enforcement, stationary patrol cars are almost always more effective than moving patrol cars. Only Shinar and Stiebel (1986) report contradictory findings. Marked police cars are generally more effective than unmarked cars (Shinar and McKnight, 1984), although the general long-term deterrence effect of unmarked cars is almost impossible to measure.

V.3.3 How much to enforce (enforcement intensity)?

Most of the enforcement experiments described in the literature are limited in terms of scope -- i.e. size and duration. The experiments have shown that enforcement operations generally

83 produce reductions in the amount of violations, with some associated halo effects, and, in some cases, reductions in accidents. However, it is unclear what levels of enforcement are needed to produce lasting changes in behaviour. Even with increased levels of enforcement, the risk of detection remains low. Hauer et al., (1980) report on an experiment with varying levels of repetition. For a single application of enforcement (of speeding), the effect seemed to vanish after three days. When the speed limit at a site was enforced for five consecutive days, the effect lasted for at least six days after the last day of enforcement. Makinen et al., (1991) report on a survey they conducted where over half the drivers questioned were stopped by a policeman over a year ago, some 10% had never been stopped and only some 9% had been stopped during the past year.

When the subjective risk of apprehension is raised sufficiently, behaviour can change in a lasting manner and enforcement levels be reduced. But, what higher levels of subjective risk of apprehension are required to create changes in driving norms? In the current public climate, and with existing resources, it seems unlikely that sufficient funding will be made available in most countries to reach such levels of enforcement. The only way to reach such levels would be with the various methods of automatic enforcement that are now becoming available.

Bjornskau and Elvik (1992) and Vaa (1997) review a number of experiments where enforcement was increased considerably. Almost all of them produced reductions in the violation rate and most could also show a decrease in the accident frequency. Table V.1 summarises the results. However, as mentioned earlier, it is unlikely that such reductions will endure once the levels of enforcement return to normal. It is also totally unclear what the base levels of enforcement were in the various studies reviewed. Elvik (1996) reports, on the basis of a meta-analysis, that lasting effects in changed behaviour (and sometimes accident reductions), are achieved when enforcement intensity is increased by at least a factor of three.

A wide range of enforcement levels exist in the various studies for which those levels could be established. Freedman and Pack (1992) report on a study of police resources in various states of the United States. They showed that in 1989 a level of 207 officers per million drivers and 1.62 officers per hundred million vehicle-miles travelled (VMT) was reached. Similar figures for Israel were quite different, with 108 officers per million drivers and 2.9 officers per 100 million VMT. Seipel (1992) calculated an intensity of one policeman per 100 km of road per shift in Germany. The corresponding figure for Israel is one policeman per 15 kilometres of road.

Standards to measure and compare police intensity do not currently exist. Such standards should be developed and compared among countries. Only then can results of various studies be compared in a meaningful way. In spite of not being able to make comparisons, some indications exist that long-term, low-intensity enforcement can generate meaningful results (Leggett 1988). Similarly, effectiveness can be improved by making enforcement more “clever/tricky”. This can generally be done by introducing an element of randomness in location and time and by varying doses of enforcement -- i.e. the number of patrol cars.

84 Table V.1 Changes in violation and accident rates associated with increases in enforcement

Reference study Factor of increase in Change in Change in enforcement violation rate accident rate Munden (1966) 6-8 -35% -25 to –28% Ekstrom, Kritz and Stromgren (1966) ca 3 -13% -21 to –37% Lund and Jorgensen (1974) ca 3 No change No change Lund, Brodersen and Jorgensen ca 5.5 -37 to -45% Not given (1977) Roop and Brackett (1980) 4-8 -15% -16 to -18% Engdahl and Nilsson (1983); Aberg 0.5-1.0 No overall ca + 11% (1983) No change change, but No change 2-3 rate of very ca - 11% 3-5 high speeds ca - 12% 5-8 was reduced ca - 19% Ross (1982), Cheshire blitz ca 9 ca - 70% -30 to - 40% Amick and Marshall (1983) 3-6 -50 to -75% ca - 40% Sali (1983) 3-4 -20 to -40% ca - 17% Salusjarvi and Makinen (1988) 2.5-3.0 60 km/h: -7% +2 to 11% 80 km/h: -25% Source: OECD.

V.3.4 Where and when to enforce?

Enforcement generally cannot be applied with the same, needed, high intensity at all locations and at all times. It therefore stands to reason that it should be more concentrated on those rural roads where accident risks are high, at times when accident frequencies are higher and to target groups which are over-involved in accidents. This leads most countries to concentrate rural road enforcement to hazardous stretches, hazardous intersections and high-risk locations. It is also easier to attract widespread public support at those sites. Enforcement is also more concentrated during high-risk periods, for instance alcohol enforcement during evenings, weekends and holidays or enforcement against fatigue during holidays.

Another approach that is not yet widely applied, but is gaining support, is to apply a certain (generally small) proportion of enforcement resources in a random fashion. Leggett (1988) reports good results with an application of 15% of the enforcement hours in such a manner.

Halo effects

Enforcement effectiveness over time and distance has been assessed in many speed-enforcement studies. The effect of the implementation effort in terms of distance from the enforcement site is generally termed the distance halo effect, whereas the effect over time is termed the time halo effect. Several authors have addressed this issue. (Pedersen-Handrahan, 1991 who cites Hauer, 1980; Leggett, 1988; Riedel et al., 1988; Bjornskau and Elvik, 1992; and Ostvik, 1989.) In spite of the extensive study, little agreement is found, as shown in the findings of Leggett (1988) and described in Table V.2. The distance halo effect ranges from 2.5 km to 22.4 km (Vaa, 1997). The smallest effect is for moving patrols while the greatest effect is for random scheduling with stationary patrols. Time halo effects range from zero (Riedel et al., 1988) through a few weeks (Hauer, 1982; Ostvik, 1989) and up to a year for intensive seat-belt enforcement (Goldenbeld, 1993). A longer-lasting effect is achieved with repeated enforcement (Goldenbeld, 1993).

85 The enforcement margin

A significant issue related to enforcement is the margin tolerated before tickets are issued. This applies in particular to speed enforcement. The police are generally inclined to tolerate a 10-20% margin on driving above the speed limit before ticketing drivers. This is justified by the inaccuracies of speedometers and by the unwillingness of the police to appear petty in court. Another point of some importance relates to the actual speeds enforced. There is a tendency on the part of the police to concentrate efforts on the more extreme speeds. These, however, are generally infrequent and are irrelevant to the large majority of drivers. It is therefore preferred from both benefit-cost and efficiency considerations to commence enforcement immediately above a certain margin. Punishment can be made progressive with the extent of the violation.

V.4 Enforcement practices

As previously mentioned, there are various types of behaviour that can be enforced and the determination of what to enforce is dependent on the authorities, most often the police, who are making the decision. Once it is determined to enforce a certain behaviour, there are various enforcement practices that can be applied for each type of behaviour. The following sections describe the practices that can be used for the principle behaviour types as well as presenting general material on automatic enforcement.

86 Table V.2 Comparisons of rural enforcement programmes (data included where available)

Study/programme Ref. Deployment method Length of Vehicle-hours Km enforced per Length of speed Suppression of patrolled routes of patrol per vehicle-hour per suppression halo mean speed day day observed per achieved per vehicle vehicle Emphasis patrol (Illinois) (a) Moving patrol 5.3 (b) Moving patrol nil Stationary patrol 9.6 km (c) Moving patrol nil (d) Stationary patrol 60 m- 6.4 km (e) Randomised scheduling, stationary 27 km .86 31.5 22.4 km 4.8 km/h vehicle (f) Stationary patrol 2.5 km Concentrated Traffic (g) Routine enforcement at high accident 2 226 km 266 8.4 Enforcement Programme times and places (Illinois) Project Increased Traffic (h) Routine enforcement at high accident 1 509 km 128 11.8 1.8 km/h Law Enforcement times and places (Texas) Offence Deterrence Randomised scheduling at high 43 km 2.02 21.2 3.6 km/h Programme (Tasmania) accident period of day; stationary vehicles

Study/programme Ref. Deployment method Relative accident Previous Benefit: cost Associated Duration of reduction (admitted accident rate ratio publicity campaign programme to hospital and fatal) per km per year evaluated Emphasis patrol (Illinois) (a) Moving patrol “fewer crashes” 6 months (b) Moving patrol Stationary patrol (c) Moving patrol (d) Stationary patrol (e) Randomised scheduling, stationary 4 weeks vehicle (f) Stationary patrol Concentrated Traffic (g) Routine enforcement at high-accident 27% 8.1 25:1 yes 6 months Enforcement Programme times and places (Illinois) Project Increased Traffic (h) Routine enforcement at high-accident 6% 1.7 3.4:1 yes 1 year Law Enforcement times and places (Texas) Offence Deterrence Randomised scheduling at 58% 0.58 4:1 no 2 years Programme (Tasmania) high-accident period of day; stationary vehicles Source: Leggett (1988). (a): Huffman et al., (1961), cited in Armour (1984) (d): Marata and Kobayashi (1972), cited in Armour (1984) (g): O'Brien (1980) (b): Smith (1962), cited in Edwards and Brackett (1978) (e): Edwards and Brackett (1978) (h): Roope and Brackett (1980) (c): Council (1970) (f): Hauer, Ahlin and Bowser (1980) V.4.1 Speed limits

Speed limits are generally set according to the type of road and geometry, type of development (urban/rural), category of activity (near schools) and type of vehicle. Rarely are speed limits differentiated according to weather (lower limits during rain or winter -- such as in Finland or France). Speed limits are usually not very flexible and, once set, they apply rigidly to varying traffic conditions. This is one of their major drawbacks. Their big advantage is that they are easily enforceable. Police have equipment, partly automatic, which can measure and record speeds above the limit. The offence is absolute in the sense that no further proof is required as to the hazard associated with the particular speed measured. A variety of speed measurement techniques and equipment have been established. These include: timing a vehicle over a known distance (Vascar), radar, radar in combination with photography as evidence and, more recently, laser-radar.

In both the United States and Australia, the most frequently used technique is the hand-held or vehicle-mounted “along-the-road” radar. It is a relatively simple and cheap device which enables police to detect vehicles approaching or retreating and has a range of up to several kilometres. Its disadvantages include its reduced applicability in dense traffic -- generally not a problem in the rural setting -- and various restrictions such as its use near high-voltage lines and metal-beam barriers. In Europe, the most commonly used equipment is the photo-radar. The radar activates a camera which takes a picture of the violating vehicle. It measures vehicles across the road -- generally at an angle of about 15-25 degrees -- though it is limited in range to 35-50 metres. Its advantages are its accuracy, recording of evidence and applicability in dense traffic. Its main disadvantage is that it is relatively expensive, thus making it unlikely that every traffic police car can be equipped with such a device.

From these relative advantages-disadvantages, the automatic, or semi-automatic mode of speed measurement evolved. This method will be described in more detail in V.4.6. The advantage of automatic speed enforcement is in the higher efficiency of the expensive equipment, the improved effectiveness in recording violations and the possibility of unmanned use.

A concept for improving speed-enforcement practices that is gaining support in many countries -- including the Nordic countries, the Netherlands, Australia, the United States and Canada -- is that of a speed-management programme (Knowles et al., 1997 and TRB, 1998). This comprises:

• improved data gathering and analysis related to speeding and its consequences -- i.e. risk analysis; • rational speed zoning; • development of speed-management and enforcement technologies; • development of effective enforcement methods, strategies and programmes; and • informational and educational approaches on speeding and its consequences for the public, law enforcement agencies, engineers and the judiciary.

A comprehensive approach such as this can clearly improve speed-enforcement practices and applications which will, in turn, lead to reductions in accidents on rural roads. In some countries, these integrated approaches act as a general deterrent by making speeding socially unacceptable.

88 Box V.2 The importance of speed enforcement

The enforcement of speed limits is one of the more controversial issues in traffic police enforcement. In many countries, exceeding the speed limit is a widely observed phenomenon. It is not uncommon on many types of road to find average travel speeds which exceed the speed limit (see Chapter II). This is certainly the case on motorways in many countries, on high-quality rural roads and on divided highways in towns, but it is also true on other types of rural roads. It is unrealistic to expect police to rigidly enforce behaviour which the majority (or a large proportion of the public) perpetrate. After all, traffic law enforcement is carried out against a background of accepted social norms and if police enforcement diverts from this social norm by too wide a margin, the public resistance it creates through the press and the political process will soon make traffic police work difficult. It might be possible to adjust the social norms slowly over time, as can be seen with drinking-and-driving in some countries, but meanwhile, traffic police enforcement operates within the margins of public support. It is, in part, because of this public resistance that in many countries speed-limit enforcement does not constitute a high proportion of traffic police duties.

Another reason for the resistance to massively enforce speed limits is the difficulty in proving “exceeding the speed limit” when a driver is involved in an accident. Generally, behaviour such as incorrectly changing lanes, not giving way at an intersection or at a pedestrian crossing, or following too closely are offences that are easier to prove in court. It is, however, quite clear that in many accidents, a lower speed adopted by at least one of the drivers could have prevented the accident or at least mitigated its results. Lower approach speeds increase a driver’s willingness to reduce speed or give way at a pedestrian crossing, at a major-minor road junction or to stop at the change of the lights at a traffic signal. In all such accidents, the “cause of accident” as reported by the police is generally not given as “high speed”. It follows that speed as an accident cause is generally under-reported.

Research findings over the past years have shown many instances of reductions in accidents and fatalities accompanied by relatively small reductions in speed (a few kilometres per hour on average) (Finch et al., 1994). In addition, more direct research on the effect of traffic-speed enforcement has generally documented accident decreases associated with reductions in travel speed. This appears to indicate that from the viewpoint of effectiveness in police enforcement and in resource allocation, it may well be justified to put considerable emphasis on speed and speed-limit enforcement.

V.4.2 Intersections

As shown in Chapter II (Table II.7), a considerable proportion of accidents on rural roads occur at intersections. Some behaviours which can be enforced by police include stopping at stop signs, giving way and not running red lights. Not much is known of the effectiveness of such enforcement tactics. Red-light cameras are becoming more widespread and to a large extent are automatic. Also, increased attention to speed enforcement can also lead to improved intersection behaviour by lowering approach speeds. This is an area where engineering and perhaps advanced technology can probably be much more effective than enforcement.

V.4.3 Alcohol-impaired drivers

This is not an area of activity mainly limited to rural roads. Techniques for the enforcement of drink-driving, backed up by well-targeted media campaigns, are well established in most countries. Almost all countries have alcohol limits, though they vary in level. Some countries allow police to apply random breath tests, whereas many other countries do not. In particular, Australia (Victoria, New South Wales and South Australia), seems to have achieved good results with very high levels of enforcing the alcohol limits. Levels of enforcement there are generally much higher than in most other countries (Hakkert, 1996 and Cameron et al., 1994).

89 V.4.4 The use of safety devices

After initial high levels of enforcement during the introduction of safety devices (mostly seat-belts in the front seats of cars) in the initial stages of legislation, seat-belt enforcement does not seem to be an area of high priority of police activity. In many countries, high levels of compliance have been achieved for front seat-belt use mostly on motorways (80-90%). However, levels of seat-belt use in urban areas are generally much lower, and rear seat-belts are infrequently used in many countries. Child safety seat regulations also exist in most countries, but again, such laws are not rigidly enforced. More publicity, together with increased enforcement, could improve the public’s use of all of these well-proven safety devices. Clearly, for the three principal accident types that occur in the rural setting, the use of safety devices can help to mitigate the seriousness of the accident.

V.4.5 Fatigue

The relationship between fatigue and accident involvement is not very well known and documented, but has merited much professional attention over the past years. Special conferences are held on the subject. Fatigue-related driving is a difficult issue to enforce for the general public because of the large number of drivers involved and their heterogeneous trip patterns. It is therefore mostly limited to holiday times when special enforcement and publicity campaigns are conducted. For professional drivers, however, this issue can also be approached through the enforcement of working hours and through the application of technology (tachographs). With the introduction of technology, this issue could also be monitored remotely (Brown 1997). In some countries, fatigue-management schemes that couple driving hours with driver health and fitness to comprehensively address fatigue are becoming more common.

V.4.6 Automatic enforcement

As discussed in the previous sections of this report, the ultimate goal of police enforcement is to achieve normative changes in the public’s behaviour to come into closer agreement with the laws and regulations. Only in this manner is it likely that enforcement can be effective within reasonable bounds of manpower and resources.

Such normative changes are unlikely to be achieved with the levels and methods of enforcement presently used in most countries. Whereas some changes, at times drastic changes, can be achieved in road-user behaviour, these changes are generally limited in time and distance. Likewise, the user behaviour changes do not result in a lasting decline in the number of road accidents associated with these types of behaviour. The explanation seems to lie in the fact that normative changes are not achieved because the variations are not internalised. It is believed that a considerable increase in risk of apprehension is required in order to achieve the desired effect.

As described in this section, such an effect might be achieved with the aid of a variety of methods of automatic enforcement. This type of enforcement can be applied on a massive scale. The technologies have been developed to make this kind of enforcement possible and feasible. In many countries, the law has been adapted to make automatic enforcement possible, and public opinion, although not always very favourable, seems to agree with a gradual introduction of this type of enforcement.

90 Red-light cameras

The use of automatic cameras at signalised intersections started in Europe in the early 1970s. The cameras are generally activated one second after the onset of the red signal. A vehicle crossing on red is captured twice, with a one-second interval between photographs to document its progress.

Photo-radar

A newer application of automatic enforcement is the use of cameras in combination with radar to catch drivers travelling at a certain margin above the speed limit. Although one of the first applications refers to the early 1980s in Germany (Lamm and Kloeckner 1984), most applications date to the late 1980s, some with considerable effectiveness. Cameras are set at the roadside, in either fixed or moveable installations, and are triggered by radars set above the speed limit. The first installation mentioned above controlled a dangerous motorway site in Germany. Speeds dropped by some 20 km/h and accidents were reduced by about 75%. In one year -- 1982 -- at this one site, some 70 000 drivers were fined.

Additional experiments conducted in a number of cities in the United States generally produced favourable results. In Galveston County, Texas, during one year, some 5 000 citations were issued (Blackburn et al., 1989). In Arizona (1987), during one year, some 10 000 citations were issued (Fitzpatrick 1992). Some applications in the United States have introduced another novelty by handing over the entire process of filming, developing and ticketing to a private firm. In most cases reported, the majority of drivers pay the fines (some 50-70%). The average speed reductions achieved are modest -- on the order of a few km/h -- but the highest speeds are reduced by a much larger amount. It has been documented elsewhere that modest reductions in average speeds can have considerable safety benefits (Nilsson 1991). The accident reduction benefits of speed enforcement cameras are now proven. The United Kingdom Home Office commissioned a cost-benefit analysis of red-light and speed-enforcement cameras (Police Research Group, 1996). The report showed a 28% reduction in accidents at camera sites. There is now clearly established public support for this type of enforcement (Freedman et al., 1992; Andersson 1990).

Some further innovative uses of automatic enforcement have been reported (Blackburn, et al., 1989), including unattended radar and portable radar associated with warning signs indicating driver’s speeds. No quantitative results of effects are reported.

Institutional bottlenecks

Once automatic enforcement becomes more widespread, a number of issues arise immediately and should be dealt with. Most importantly, it is generally the case that the vehicle is not stopped. Therefore, the vehicle owner is the violator unless he can prove otherwise. This is now legally accepted in various countries, including the Netherlands, Germany, United Kingdom, Israel, and some jurisdictions in the United States and Canada. The offence is, to some extent, decriminalised, and becomes more fiscal in nature. In some countries, however, such offences can also incur penalty points that, when accumulated, can lead to licence revocation. Other bottlenecks that should be dealt with to make this type of enforcement effective are a shortening between the date of offence and the issuing of a ticket, the legal process if a driver chooses to go to court and the financial cost that is incurred by the police for processing these offences. In addition, the deployment of many technologies such as this is accompanied by serious privacy issues in some countries.

91 V.5 Management, priority setting and funding

As appears from the foregoing there are several possibilities to improve road safety on rural roads with police enforcement (to be understood in a broad sense). It must, however, be noted that there are different reasons why enforcement does not take place, or not to the full extent required. Enforcement could therefore be making a greater contribution to the improvement of road safety than is currently the case in many rural jurisdictions.

It is clear that enforcement must compete with other activities -- crime, drugs, etc. -- within the existing (sometimes shrinking) budgets of the police and judicial organisations. Experience in different OECD Member countries shows that when choices have to be made, traffic law enforcement is given a relatively low priority. In this connection, it is also a fact that within the police organisation enforcement is judged by management and executive police officers as not being a very attractive activity -- i.e. “enforcement does not score”. Police culture is another hindrance because it is primarily aimed at tracking down and identifying criminal violations of the law, determining the question of guilt and subsequently penalising offenders. In contrast, traffic enforcement is concerned above all with influencing road-user behaviour. Prevention is next to repression as an essential feature of effective enforcement, but this idea does not always fit well in existing police culture. Also, co-operation with mass media is in many instances a new concept for police and judiciary.

Another marginal note is related to the learning capacity of police and judicial organisations. It is, at most, incidental practice to document police activities in such a way that relevant management information (input, throughput and output indicators) is available on a basis upon which enforcement can be optimised. It is more the rule than the exception that enforcement in the form of temporary projects is organised where the investments in terms of time and money are so high that, should the enforcement be successful, there is no possibility to implement the successful enforcement on a larger scale and over a longer period of time.

Although present police practice and the associated good examples are not exhaustively characterised in the above-mentioned cases, they do serve to illustrate that effective and efficient enforcement (again in a broader sense) does not rely solely on essential knowledge, but also requires management, setting (higher) priorities and financing for enforcement and public relations. In this framework, a few suggestions which might address the above-mentioned problems can be offered.

First, it would be considered “good practice” if the police based enforcement strategies on sound accident analyses and available, scientifically supported insight on enforcement effectiveness. Such a rational approach should be standard in police management. The information which becomes available through this approach can be used as feedback to the police personnel to increase their motivation and performance. Such an approach could also be beneficial in ex post evaluation of the effectiveness of targeted campaigns.

Second, if the capacity of the police and judiciary is a bottleneck for adequate enforcement, some minor offences could be settled by means of judicial administration rather than by penalty, if this fits within the legal system. A second possibility in this regard would be to have portions of the work -- i.e. those activities where no contact with the road user is necessary -- carried out by non-executive employees. Another possibility would be to make a well-considered choice between more frequent use of modern technology -- for example, speed cameras and red-light cameras -- and less use of police officers. If the use of personnel in the execution of special enforcement tasks has a low priority, then these tasks should be automated to the extent possible. Also, fines that are not collected on the spot but are instead issued electronically and sent to vehicle owners at a later date should be carefully managed. It is also

92 recommended that these activities be seen as a chain of events where vulnerable links should be adverted. A final option to mention here is the possibility of integrating different measures in one campaign in order to decrease the total costs for the police force (Mathyssen, 1992). A third possibility for overcoming the enforcement hindrance caused by a lack of manpower and/or financing is to use some or all of the proceeds of traffic fines to finance the costs of enforcement. In this respect, the (pre)financing of investments in equipment or the financing of the exploitation costs of the enforcement are examples where this could be especially useful. For example, in Queensland, Australia, the introduction of speed cameras was successful on the basis that revenues generated through their use were dedicated to road safety (not just enforcement). Although these considerations can only be mentioned here briefly and cannot necessarily be applied everywhere, they give a line of reasoning to overcome the “traditional” problems of lack of enforcement resources and manpower. They could also help to prevent traffic law enforcement from being at the bottom of the priority list while, at the same time, maintaining a balanced budget. Furthermore, it is important to create conditions in which traffic law enforcement is not vulnerable to day-to-day disturbances. Special attention from upper levels of police management, well-trained police officers equipped with modern technology and good co-operation between the police and judiciary are essential factors for effective and efficient traffic law enforcement.

V.6 Conclusion

Police enforcement is effective in reducing crashes. It is also a symbol to show (to citizens in general) that road safety is important. If police are not present, it sends a message that traffic violations are less important than other types of crimes and misdemeanours, though they could be equally or more dangerous for society. Effective enforcement can serve as a general deterrent factor that will change behaviour. However, for the changes to be long-lasting -- i.e. lead to a change of norms -- they must be coupled with other firm actions including appropriate punishment and sufficient driver training. In addition, coupling publicity campaigns with targeted enforcement is gaining ground and has proven to increase the enforcement effects which can also contribute to a change in driving norms. A number of general conclusions can be brought forward on enforcement strategies and tactics that are relevant to the rural road setting. For instance, repeated enforcement creates longer halo effects, in terms of either time or distance. This is to be contrasted with the “blitz” campaigns that are undertaken more often than permanent measures. By introducing a random enforcement element, enforcement effectiveness can also be increased and longer halo effects will be produced. Finally, because of the high cost associated with the level of manned enforcement required to make a difference on the extensive rural road networks, automated enforcement technologies targeted to address the principal rural road accident types may offer a valid approach to help reduce the number and severity of these accidents. Traffic enforcement levels on rural roads are generally not consistent with the accident rates on these roads, especially given the fact that nearly 60% of road accident fatalities in OECD countries occur on rural roads. This situation is brought about by a lack of funds for rural road traffic enforcement. These funding gaps occur either due to a low priority for traffic enforcement in budgeting processes or because rural roads lose when competing with urban areas for limited traffic safety funds. It is therefore suggested that funds be earmarked for rural road safety to ensure that these important safety problems are addressed to the fullest extent possible. One option for obtaining earmarked funds is to dedicate the revenues from traffic fines to police enforcement activities. It is further recommended that enforcement strategies -- including the application of automatic enforcement technologies -- be put in place and that enforcement management systems be used to ensure the rational and appropriate implementation of these strategies.

93 BIBLIOGRAPHY

ANDERSSON, G. (1990), Speed Limits and Tolerances of Enforcement, Severity of Sanctions and Intensity of Surveillance, Proceedings of the Symposium on Driving Behaviour in a Social Context, T. Benjamin (Ed.), Publisher Paradigme, Paris.

ANON (1985), Staysafe 5 - Appropriate Strategies for Police Traffic Law Enforcement, Parliament of N.S.W., Joint Standing Committee on Road Safety, N.S.W., Australia.

ARMOUR, M (1984), The Effect of Police Presence on Urban Driving Speeds, Australian Road Research 14(3).

BJORNSKAU, T. and ELVIK, R. (1992), “Can Road Traffic Law Enforcement Permanently Reduce the Number of Accidents?”, Accident Analysis and Prevention, Vol. 24, No. 5, pp. 507-520.

BLACKBURN, R.R., MORAN, R. and GLAUZ, W.D (1989), Update of Enforcement Technology and Speed Measurement Devices, U.S. Dept. of Transportation, National Highway Traffic Safety Administration, Midwest Research Inst., Kansas City, Missouri.

BROWN, I.D (1997), “Prospects for Technological Countermeasures against Driver Fatigue”, Accident Analysis and Prevention, Vol. 29, No. 4.

CAMERON, M., NEWSTEAD, S.V. and VULCAN, P. (1994), “Analysis of Reduction in Victorian Road Casualties, 1989 to 1992”, Proceedings of the 17th Australian Research Board Conference, Gold coast, Queensland, Australian Road Research Board, Melbourne.

CORBETT, C., SIMON, F. and O’CONNELL M. (1998), “The Deterrence of High Speed Driving: A Criminological Perspective”, TRL Report 296, Transport Research Laboratory, Crowthorne, Berkshire.

ELVIK, R (1997), The Effects on Accidents of Automatic Speed Enforcement in Norway, TRB Paper 970118, TRB, Washington, D.C.

FINCH, D., KOMPFNER, P., LOCKWOOD, C. and MAYCOCK, G. (1994), “Speed, Speed Limits and Accidents”, TRL Report 58, Transport Research Laboratory, Crowthorne, Berkshire.

FITZPATRICK, K. (1992), A Review of Automated Enforcement, Compendium of Papers, Institute of Transportation Engineers.

FREEDMAN, M. and PACK, N.N (1992), Police Enforcement Resources in Relation to Need: Changes during 1978-89, Insurance Institute for Highway Safety, Washington, D.C.

GOLDENBELD, C. (1993), Compliance with Traffic Regulations in the Netherlands, Report in Dutch (Handhaving van verkeersregels in Nederland), SWOV, The Netherlands.

94 HAKKERT, A.S. and MCGANN, A (1996), Road Safety in Australian States, Research Report 96, Australian Road Research Board, Vic.

HAUER, E., AHLIN, F.J. and BOWSER, J.S (1980), Speed Enforcement and Speed Choice, Transport Canada, Publication 80-02, Canada.

KNOWLES, V., PERSAUD, B., PARKER, M. and WILDE, G (1997), Safety, Speed and Speed Management -- A Canadian Review, Transport Canada Research Report, ASF 3261-280, Canada.

LAMM, R. and KLOECKNER, J.H (1984), Increase of Traffic Safety by Surveillance of Speed Limits with Automatic Radar Devices on a Dangerous Section of a German Autobahn: A Long-term Investigation, TRB Record No. 974, TRB, 1984, pp. 8-16, Washington, D.C.

LEGGETT, L.M.W (1988), “The Effect on Accident Occurrence of Long-term Low-intensity Police Enforcement”, Proceedings, 14th ARRB Conference, Part 4.

MAKINEN, T. and SYVANEN, M. (1990), “Police Surveillance Strategies”, Proceedings of the Symposium on Driving Behaviour in a Social Context, T. Benjamin (Ed.), Publisher Paradigme, Paris.

MAKINEN, T., BEILINSON, L. and SALUSJARVI, M (1991), “Traffic Enforcement Strategy and Tactics”, 2nd International Conference on New Ways and Means for Improved Road Safety and Quality of Life, Tel Aviv.

MATHYSSEN, M.P.M (1992), “Integral Enforcement of Alcohol Consumption, Speed, Use of Seat Belts and Helmets”, Report R-92-19, SWOV, Institute for Road Safety Research, Leidschendam.

NILSSON, G (1990), “Speed Limits, Enforcement and Other Factors Influencing Speed”, in Koornstra and Christensen (Eds.): Enforcement and Rewarding Strategies and Effects. Proceedings of the International Road Safety Symposium, Copenhagen, September 19-21, 1990, SWOV, Leidschendam, The Netherlands.

OSTVIK, E (1989), Road Traffic Law Enforcement -- Conceptual Framework for Police Surveillance Techniques, Report No. ISBN-82-7133-646-0, TOI, Institute of Transport Economics, Oslo.

PEDERSEN-HANDRAHAN, K (1991), “An Evaluation of the Radar Camera on Traffic Speed”, Proceedings of the Canadian Multidisc Road Safety Conference VII, Vancouver.

POLICE RESEARCH GROUP (1996), Cost-Benefit Analysis of Traffic Light and Speed Cameras, Home Office, London.

RIEDEL, W., ROTHENGATTER, T. and DE BRUIN, R (1988), “Selective Enforcement of Speeding Behaviour”, in Road User Behaviour: Theory and Research, Rothengatter, de Bruin (Eds.), Van Gorcum, The Netherlands.

SEIPEL, C. (1992), Polizeiliche Uberwachung im Strassenwerkehr aus socialwissenschaftlicher sicht -- Planungstudie, Internal Report, BASt, Bergisch Gladbach.

95 SHINAR, D. and MCKNIGHT, A.J. (1984), The Effects of Enforcement and Public Information on Compliance in Human Behaviour and Traffic Safety, Evans, L. and Schwing, R.C. (Eds.), Plenum Books, New York.

SHINAR, D. and STIEBEL, J. (1986), “The Effectiveness of Stationary vs. Moving Police Vehicles in Compliance with Speed Limit”, Human Factors, 28(3), pp. 365-371.

TRANSPORTATION RESEARCH BOARD (1998), “Managing Speed: Review of Current Practice for Setting and Enforcing Speed Limits”, Special Report 254, TRB, Washington D.C.

VAA, T. (1997), “Increased Police Enforcement -- Effects on Speed”, Accident Analysis and Prevention, Vol. 29, No. 3.

ZAAL, D. (1994), Traffic Law Enforcement -- A Review of the Literature, Accident Research Centre, Monash University, Victoria, Australia.

96 Chapter VI

POTENTIAL SOLUTIONS PROVIDED BY INTELLIGENT TRANSPORT SYSTEMS

VI.1 Introduction

Technological advances in computing capabilities, communications, electronics and geo-positioning systems are all contributing to advances in Intelligent Transport Systems (ITS) that can be used to improve rural road safety. Information capabilities have been deployed in urban areas for many years to improve the operation and efficiency of the urban highway network, and information on real-time traffic conditions is available to drivers in many of these areas today. As the technology improves and as other applications are identified through research and development efforts, the expansion of ITS applications will also help to solve rural safety problems. For instance, emerging technologies and real-time information capabilities can be used to provide drivers on rural roads with general information about the weather they will encounter or location-specific information about the physical layout and condition of the road ahead. Through applications such as these, ITS, when deployed, will reduce the number of accidents, injuries and deaths in road transport, including on rural roads.

Vehicle research is producing additional on-board capabilities for the future that can contribute to other ITS applications that take advantage of interactions between the vehicle and the roadside. For instance, on-board vehicle sensors for lateral positioning, obstacle detection and tyre-road surface friction monitoring can be integrated with location-specific information to help drivers maintain safe operation of their vehicles. Eventually, control intervention systems will be developed to automatically assist drivers in hazardous situations. Someday, with sufficient research, limited vehicle control could give way to full automated control of vehicles, thus producing even higher levels of road safety.

Advances in ITS technologies that can improve road safety will be made as the uncertainties surrounding driver behaviour are eliminated. This is being accomplished through carefully designed research programmes that are underway in many countries. At the same time, if they are going to accept them, drivers must understand the capabilities and limitations of technologies that assist the driver. This understanding will develop gradually during the transition of the vehicle fleets over several decades. Additionally, resistance by certain segments of the driving population -- older drivers, for example -- may present unique barriers to the rapid and widespread deployment of ITS technologies.

This chapter discusses ITS technologies that have the potential to improve rural road safety in the future. It clearly identifies the technologies that are available now, those that are in the development phase and others that still require research. This chapter does not describe technologies associated with automatic enforcement -- i.e. red-light cameras and speed enforcement, or trauma management -- i.e. mayday systems, because those technologies are discussed in chapters V and VII, respectively. The chapter concludes with an analysis of the relative costs of certain technologies and the general timeframe in which they will be available for deployment.

97 VI.2 General safety considerations of ITS

From the foregoing text it is clear that there is reason to anticipate positive safety effects being derived from the deployment of ITS. It is, however, important to point out that an essential step in the deployment of any ITS technology is to consider what the safety impacts -- both positive and negative -- of the technology could be and to have in place the means to evaluate those safety impacts at various stages of the deployment from planning and design, to early operation and long-term performance. This is especially relevant in the rural road environment where the balance of costs and benefits of any safety measure is of high importance.

The European Transport Safety Council (ETSC, 1998) has listed a number of relevant ITS safety issues that are also applicable for the implementation of ITS on rural roads. Both positive and negative effects can be considered in relation to any of the following:

• direct in-car modification may influence driver attention, mental load and decision making -- e.g. choice of speed, etc -- all of which may influence safety; • direct influence by roadside systems will have similar safety impacts as in-vehicle, though more limited -- e.g. route changes; • indirect modification of ITS user behaviour -- i.e. behavioural adaptation -- could come about in many, largely unknown ways and may appear in the form of such actions as a change of headway in a car - following situation or a change in the expectation of the behaviour of other road users; • indirect modification of non-ITS user behaviour which may appear as non-equipped users imitate the behaviour of ITS-equipped drivers by driving closer or faster than they should; • modification of the interaction between ITS users and non-users; • modification of road-user exposure by changing travel patterns, route choice, modal choice, etc.

The ETSC points out that the challenge in considering what the positive and negative impacts will be of these and other changes arises in their evaluation. One important challenge in this regard is the fact that it is very difficult to measure the safety impact of ITS by traditional methods based on the number of crashes or injuries given that there is not a large number of ITS users on the road today. However, they suggest that because ITS deployment will be market driven, it is imperative that any introduction of a new technology or application should be accompanied by road safety monitoring. Monitoring should include both pre-deployment evaluations as well as short- and long-term evaluations of the safety performance of the technologies. Though many existing means of safety evaluation can be employed, there is a need for further research into the safety evaluation of ITS.

VI.3 Speed-control devices

In the previous chapters, speeding and inappropriate speed were shown to be the most important factors in rural road crashes. ITS can improve rural road safety with the application of systems that are intended to help drivers select appropriate speeds for the given conditions or maintain the posted speed limit. For instance, simple roadside active warning systems can alert drivers to situations where their real-time speed is not appropriate for the traffic, geometry or environmental conditions ahead and encourage them to adjust their speed. Unlike these passive systems that require a response from the

98 driver, there are other ITS technologies that are active in that they adjust or control the vehicle speed automatically. Whether active or passive, speed-control devices are important for the rural road environment because they can be particularly effective countermeasures for run-off-the-road and intersection accidents.

VI.3.1 Speed advisory systems

Several countries have utilised speed sensors to alert drivers that their speed is inappropriate for the conditions ahead and could lead to a hazardous situation if not reduced. Typically, these speed advisory systems are deployed at horizontal curves or on steep downgrades. The sensors activate flashing beacons in combination with either static or changeable message signs. Operational field tests in the United States were very effective in reducing speeds and reducing the number of run-off-the-road accidents (FHWA, 1997). It was also determined that the number of heavy vehicle roll-over accidents can be reduced by using speed advisory systems. In Australia and other countries, speed advisory signs are linked to systems that monitor for adverse weather conditions such as fog and, in some situations, are mandatory for enforcement purposes.

Another speed advisory system application is portable signs that display the vehicle approach speed. These systems are primarily for information purposes only and are intended to make drivers aware of their speed. There are no conclusive studies on the safety benefits of these speed information signs, although evidence shows significant reductions in vehicle speeds (TRL, 1996). For example, trials in the county of Norfolk (United Kingdom) using speed-activated signs on approaches to bends, junctions and villages have shown that significant speed reductions can be achieved, with corresponding reductions in accidents and casualties. With further research, these systems may prove to be a useful tool for rural road safety.

VI.3.2 Speed adapters

On-board equipment to control the speed of a vehicle has been in use for many years. In addition to maximum speed governors to limit the speed of trucks and buses, the technology now makes it possible, based on long-available cruise control devices, to equip cars as well with such devices which can be adjusted by the driver, as before, or independently of the driver through communication links with roadside systems. These can provide location-specific information needed to automatically adjust vehicle speed according to the speed limit and therefore increase safety by preventing violations. Research in the Netherlands and Sweden is underway to develop intelligent speed adaptation (ISA) by integrating location-specific, maximum speed into the vehicle’s control system. The integration is possible by providing information, or a warning if necessary, to the driver so that speeds can be adjusted either manually (voluntary) or through automatic intervention (mandatory). A field test of the application was initiated in 1998 in Tilburg. The deployment of this technology, like many other ITS applications, requires acceptance by drivers. Other countries, including Sweden and the United Kingdom, are investigating similar speed-control systems and these studies will include, of necessity, evaluation of the legal and institutional implications of speed intervention.

99 VI.3.3 Speed-control assistance

A speed-control assistance system is intended to improve safety by automatically assisting in the control of vehicle speed. This type of system is being studied and developed as a means for coping with a variety of obstacles and dangerous conditions, including the following:

• obstacles in the vehicle path; • inter-vehicle relations at intersections and on roadways; • slow-moving vehicles such as construction and maintenance equipment, as well as personal vehicles in specific conditions such as increasing traffic density; • road features such as curves; • hazardous road surface conditions such as potholes, water or ice; • weather conditions including low visibility; and • the physical condition -- i.e. fatigue, etc. -- of drivers.

The development of speed-control systems for curves would be very beneficial in rural areas as they could have the ability to calculate the curvature of the upcoming stretch of roadway and determine the appropriate speed for the vehicle to remain within its lane. If the vehicle is travelling too fast, the system warns the driver and automatically decelerates the vehicle.

A system being implemented in Japan uses a navigation system which detects the position of the vehicle and decelerates speeding vehicles as they enter curves. First, a Global Positioning System (GPS) uses satellites to detect vehicle position, while an in-vehicle system determines the travelling lane by way of sensors on the right and left wheels. This data is compared to an electronic road map (map matching) to more precisely estimate the vehicle position. The appropriate speed limit is then determined based on the radius of curvature of the roadway. Deceleration control is managed with a warning alarm that factors in the speed limit. Finally, the vehicle speed is adjusted by throttle and transmission control.

Other systems under development rely on vision-based detection using cameras to recognise travelling lanes and calculate roadway curvature. Systems that use vehicle-to-roadside communications infrastructure to calculate roadway curvature and limit vehicle speed have been demonstrated in projects in Europe and Japan.

VI.3.4 Adaptive cruise-control systems

Systems which combine cruise-control and headway-distance warning functions are generally referred to as “Adaptive Cruise Control”. When the vehicle is operating in a manner unconstrained by other vehicles, its speed remains as pre-set by the driver. When the vehicle approaches a slower vehicle in the same lane, the system controls the vehicle speed to maintain a safe following distance. This type of system was studied and developed under the Programme for European Traffic with Highest Efficiency and Unprecedented Safety (PROMETHEUS) (European Commission, 1997). A similar system was developed in the Advanced Safety Vehicle (ASV) Project in Japan (Ministry of Construction, 1996) and is currently being evaluated, and in an “Intelligent Cruise Control Field Operational Test” in the United States (ITS America, 1997).

100 VI.4 Driver/vehicle information systems

There are several ITS technologies under development that will contribute to road safety both from the perspective of driver performance as well as the driving environment. This section deals with safety-related systems that provide information to assist drivers in improving their own performance or to make better routing decisions and thereby avoid hazardous situations. Systems that are intended to mitigate the effects of accidents once they occur are also covered.

VI.4.1 Driver-management technologies

Driver-management technologies have the potential to reduce accidents due to errors caused by the physical condition of drivers. In particular, driver-monitoring systems use sensors to monitor the physical aspects of the driver -- i.e. breathing, heart rate, eye position and movements, and other aspects -- or the status of several of the vehicle control systems, most commonly the steering system, and then activate passive or active warning and control systems as necessary. Keeping drivers alert and ready to perform in the event of a needed control manoeuvre is a key ingredient in reducing run-off-the-road accidents resulting from driver fatigue. Studies in Europe have shown that accident rates can be reduced by up to 41% with simultaneous reductions in accident severity as well (European Commission, 1996-97).

Other, simpler driver-management technologies are also available that could improve rural road safety. For instance, in Australia, police cars in rural areas have single line variable message signs mounted with their emergency lights on the roof. These signs allow the driver of a following vehicle to read particular safety warning messages. To date, these signs have been used to alert drivers to the risks associated with fatigue and provide recommendations for rest breaks. When coupled with free coffee stands, this approach can lead to reductions of rural road crashes on holiday weekends and other periods of heavy travel. Other systems that are used to alert drivers to the approach of an emergency vehicle are also applicable in some rural situations. In these systems, emergency vehicles have a transmitter that sends a signal to vehicles in their path. The signal is received in the vehicle and activates a flashing light on the dashboard. Because these systems generally operate on a “line of sight”, their application may only be useful in particular rural situations.

VI.4.2 Vehicle data recorders (black boxes)

Research in Europe, including the DRIVE SAMOVAR Project, has shown that driver behaviour improvements that lead to better highway safety can result from the deployment of on-board vehicle data recorders. These data recorders -- known as “black boxes” in the airline industry -- are capable of recording objective data about vehicle operations on a continuous basis. The data can be retrieved and used for operational management/logistics -- e.g. tracking and routing dangerous goods vehicles -- or for recreating pre-crash or post-crash events. In a country such as Australia, for example, that is characterised by very remote areas, these technologies have an added value in aiding to identify the exact location of disabled vehicles. Operational field tests of vehicle data recorders in the United Kingdom, Belgium and the Netherlands have produced accident reductions estimated to be in the range of 20% (SWOV, 1997). In the United Kingdom, fleet operators are eligible for as much as a 15% reduction in their insurance premium when journey recorders are installed in their vehicles (TRL, 1996). While vehicle data recorders are not a specific countermeasure, the widespread deployment of ITS technologies that improve driver performance will produce safer overall road systems.

101 VI.4.3 Navigation systems

Navigation systems are currently available as an original equipment option in many new vehicles and in several after-market products, including notebook PC versions. They use GPS for position and navigation information. These systems permit drivers to know where they are and to have confidence about their routing decisions. By eliminating a source of driver indecision which contributes to driver error, some accidents can be avoided. To improve safety, systems are being developed that use alternative forms of output -- e.g. machine voice -- that require less attention from the driver than those systems that rely on visual outputs alone. As an additional step to ensure that safety is not compromised, other systems require the vehicle to be stopped in order to engage certain information service functions.

Most navigation units provide real-time position information to the driver and are capable of turn-by-turn routing directions in unfamiliar areas. These units also have the capability to include a listing of available driver services such as fuel, food, accommodation and emergency aid.

Portable in-vehicle message systems are in the field development stage in the United States. These small, hand-sized units are capable of informing the driver of conditions ahead through the activation of one or more pre-programmed messages. Originally designed for application in work zones, these devices will be capable of handling additional information and warning message sets in the future (FHWA, 1997).

VI.4.4 Weather information

Information on both real-time weather conditions and anticipated conditions will better prepare drivers for expected road and driving conditions. With this information, drivers can avoid hazardous conditions by planning their routes better and/or modifying their schedules. Weather information systems have already been deployed in various parts of Europe, Japan, Australia and the United States to provide for the safety and security of motorists (ITS America, 1997).

Weather advisory systems range from flashing beacons with static message signs to changeable message signs and electronic information panels with weather conditions and alerts. In some cases, the weather information is generated by location-specific weather sensors and is then used by road agencies to physically close road segments in order to reduce the danger to road users. Though winter is the most likely time of year when weather conditions will force road closures, several states in Europe and the United States also have systems that provide fog detection and warning where these conditions can be severe (FHWA, 1997). In Australia, fog detection and warning systems are linked to speed signs to reduce travels speeds during periods of adverse conditions.

Pre-trip weather information is available through many existing sources that publish or display weather data and forecasts. Some traveller-specific systems that provide this information include Internet real-time conditions and forecasts; telephone dial-up services; traveller information terminals, sometimes known as “kiosks”; roadside changeable message signs; and roadside advisory radio broadcasts using standard automobile radios. These systems are designed to give accurate information to drivers about expected weather and road conditions. For example, several of these technologies were successfully integrated and deployed for the benefit of the visitors and participants in the 1998 Winter Olympic Games in Nagano, Japan (Ministry of Construction, 1996). This example and others have shown that by providing better information to drivers, fewer accidents are likely to occur. However, it should be noted that because these systems have only recently been adopted there is little scientific evidence to support this, but considerable anecdotal evidence has been reported that indicates the success of these systems.

102 On-board vehicule-based systems have been developed to accurately monitor pavement surface temperature to assist road agencies in the application of anti-icing and de-icing chemicals. Accurate data on surface temperature at the system level could also be useful for travellers, especially for those in specific locations where accident problems caused by ice are known. Further developments are necessary in order to deploy this technology on standard passenger vehicles.

VI.4.5 Accident mitigation

On-board vehicle technologies are available to help mitigate the effects of a crash on accident victims. Research on “Smart Belt” and “Smart Air Bag” technologies is currently being conducted and should lead to the development of product enhancements that will assist in the timing and tensioning of seat-belts and the deployment of air bags during a crash sequence. These advances will help to minimise injuries and deaths in accidents, particularly the high-impact crashes which characterise many rural road situations.

VI.4.6 Crash avoidance

Some estimates suggest that driver error is a contributing factor in 90% of all crashes. Technologies designed to provide timely information to drivers through on-board sensors can help to contribute to improvements in safety on rural roads because they can reduce the perception and reaction time necessary for drivers to act in response to a potential hazard. The general heading of crash avoidance systems includes all technologies that use sensors and information systems to assist drivers. These systems either provide a warning to take some action (passive) or initiate direct and automatic intervention of an actuator to implement corrective actions (active). Passive warning systems are commercially available today. Radar-based systems and closed-circuit cameras are the most commonly used applications. The radar systems are typically employed for forward-collision avoidance and to monitor blind spots. Closed-circuit cameras can be used to detect obstacles and hazards in the vicinity of the vehicle.

In the United States, the National Highway Traffic Safety Administration estimates that three selected collision avoidance systems addressing rear-end, lane change, and single vehicle roadway departure accidents have the potential to eliminate 1.2 million crashes annually.

Additional research into promising collision avoidance technologies is underway in the United States (Intelligent Vehicle Initiative), Japan (Advanced Safety Vehicle or ASV) (Ministry of Construction, 1996) and Europe (DRIVE II, Transport Telematics Applications Programme) (European Commission, 1997). Future systems are expected to go beyond passive warning to provide active assistance through control intervention in hazardous situations.

VI.4.7 Lateral guidance systems -- a future application

The concept for lateral guidance systems began in the Automated Highway System (AHS) Research and Programmes effort and is now being further advanced through initiatives in the United States, Japan and Europe. The testing and evaluation activities taking place in these three regions are focused on a number of lateral guidance systems, including magnetic nails, vision-based systems and radar systems using specially prepared pavement marking tape.

103 The Partners for Advanced Transit and Highways (PATH) project in the United States has successfully tested vehicle location feedback control and corrective control through the use of magnetic nails. The corrective control is accomplished by using headway data from a combination of magnetic nails to estimate headway errors. The Carnegie Melon University in the United States has led the development of the so-called Navigation Laboratory (NAVLAB) which is an experimental vehicle with an automatic steering system controlled by the application of visual processing technology for lane recognition. The system was successfully used for 97% of a 5 000 km running test from Washington, D.C. to San Diego, California. Other lane-keeping systems have been developed by the private sector, through a running demonstration in Europe under the PROMETHEUS project, and in the United States as part of a initiative called “Demo 97”(ITS America, 1997). In Japan, development of lane-keeping systems has been promoted in the AHS and ASV projects and was successfully demonstrated in 1996 with 11 automated vehicles on an unopened section of the Joshinetsu Expressway (Ministry of Construction, 1996).

When they are widely available, lateral guidance systems will provide opportunities to help keep vehicles from leaving the roadway and crushing into other vehicles or roadside obstacles. The possible future applications of these technologies to improve vehicle guidance on rural roads include: narrow roadways; hazardous single curves or combination of curves; and times and locations of low visibility due to fog, rain or snow.

VI.4.8 Vision-enhancement systems -- a future application

Several countries are currently conducting research and testing to develop systems that can enhance night driving. One of the most promising technologies that is currently under evaluation in both Europe and the United States is ultraviolet headlamps. Ultraviolet headlamps have the potential to improve road safety by increasing the visibility of roadway pavement markings at longer distances and, as an additional benefit, providing a higher level of pedestrian safety. Other vision-enhancement systems that are under development include those that rely on supplemental cameras and infrared technologies. An interesting aside of enhanced-vision systems may be their use to address the safety issue related to ageing societies. The results of surveys in Europe estimate that 60-70% of elderly drivers would be willing and able to drive more with vision-enhancement systems (TRL, 1996). This may provide some relief to elderly citizens living in rural areas where public transport is not available.

VI.5 Infrastructure-based applications for co-operative systems

Systems that link information provided by roadway sensors to in-vehicle warning and control systems are called co-operative systems. In today’s highway environment, roadway sensors actuate roadside changeable message signs and/or flashing warning signals. The most traditional of these systems involves railroad highway grade crossings where approaching trains are detected automatically and flashing lights and protective gates are activated until the train passes the crossing. Other warning systems are used to provide improved levels of safety and security on rural road systems. Typically, these systems use sensors to activate roadside warning signs or devices. In the future, truly co-operative systems will be in use that allow in-vehicle information systems to be actuated by roadside transponders or system-level communications in order to alert drivers to special road conditions and to potentially adjust the vehicle-control systems.

104 VI.5.1 Intersection approach warning

Similar in application to the speed-advisory systems, infrastructure sensors such as inductive loops can be used to identify vehicles at the approaches to rural intersections. These sensors can activate warning beacons in combination with static signs to alert drivers of vehicles approaching from other directions. These technologies will be especially useful when sight distance is restricted or when intersection approach speeds are high. These systems could be particularly effective at high-accident locations.

VI.5.2 Guide-light system

A guide-light system using an LED display has been installed in Japan at the outer part of a curved section of rural highway where visibility is poor. The system is designed to display green lights to notify drivers of the existence of a curved section ahead. However, red lights flash when there are oncoming vehicles which the driver cannot see from a distance. The guide-light system was first introduced in Japan in the mid-1990s, and the number of installations is increasing (Ministry of Construction, 1996). This is a system application that is particularly well suited to the rural environment, especially given the nature of the safety problems.

VI.6 Human factor considerations

An important consideration in the development and deployment of vehicle-based driver information and assistance systems concerns how drivers will react and respond to this equipment. In general, the reliability of the systems as they become more aligned with control functions must be assured if drivers are going to allow them to be used to their full potential. More specifically, the level at which the driver understands and accepts the equipment is critical to the success of ITS applications in rural safety. Some of the questions that need to be answered as research in this area continues centre on how drivers will modify their behaviour in relation to risk tolerance. This is especially important because the presence of driver assistance systems that could be perceived as extensions of human capabilities could increase risk tolerance and therefore reduce safety. Human factor research -- sometimes referred to as driver/vehicle interface or human/machine interface -- that addresses these types of issues is underway in Europe (TRL, 1996), Japan (Ministry of Construction, 1996) and the United States (ITS America, 1997).

VI.7 ITS in perspective

Evolving ITS technologies that hold promise for future improvements in road safety will continue to require co-operation between the public and private sectors. Rapid advances in technology will continue to breed new opportunities in highway and vehicle safety that address the most common rural crash problems. As previously stated, autonomous vehicle systems will evolve from simple warning systems to intervention and control, including fully co-operative vehicle/highway systems. Table VI.1 contains selected ITS technologies, their applications and general statements about their relative costs and time to deployment. Though preliminary assessments of the relative costs and benefits of ITS technologies have been performed in Europe, for the United States and Japan more detailed analysis will be necessary as additional research and development is conducted.

105 Table VI.1 Selected ITS technologies, the timeframe in which they will be ready for deployment and their relative costs

Technology Accident type Deployment Relative Reduced Timeframe1 Cost Speed control - Advisory Run-off the road Near Low Intersection - Adaptive cruise Run-off the road Near Low Control - Roadside control Intersection Mid Medium Driver monitoring Run-off the road Near Low Intersection approach Intersection Near Low warning Guide-light warning Run-off the road Near Low Head-on Weather information Run-off the road Near Low- medium systems Intersection Crash avoidance All Near-mid Low-medium Lateral guidance - Warning Run-off the road Near Low-medium - Control Run-off the road Mid-long Medium-high Accident mitigation systems - Smart Belts/Bags All Near-mid Low Vehicle data recorders All Near-mid Low

1. Deployment Timeframe - Near = Present - 3 yrs; Mid=3-7 yrs; Long = 8+ yrs.

Source: FHWA.

VI.8 Future needs

Realising the full potential of ITS solutions to rural safety problems depends upon a number of evolving issues. On-board vehicle safety systems will evolve from autonomous single systems into fully integrated vehicle-highway systems. As these systems begin to appear in vehicle fleets, road safety will begin to improve, but major research is needed to address: the costs (and benefits) of these systems; various technical issues; human-machine interface questions; and institutional, legal and political constraints. Future consideration of the driver and the roadway as a single system could lead to rapid breakthroughs in improving road safety. The public and private sectors will both benefit from these initiatives.

Building upon the success of early European research in the DRIVE and PROMETHEUS projects, current research programmes such as the Automated Highway System work in Japan and the United States, along with the Crash Avoidance Research programme in the United States, will deliver results that are essential for the continuing deployment of ITS technologies. In Europe, projects such as AC-ASSIST, CHAUFFEUR and SAVE, will develop future applications to improve road safety. In the United States, the Intelligent Vehicle Initiative will be the focal point for highway/vehicle system research to improve safety. In Japan, a programme of research called the Advanced Vehicle Control and Safety Systems is striving for similar road safety goals. All of these aggressive research programmes provide the potential for significant improvement in highway safety in the years to come.

106 Transport officials with responsibility for rural road programmes will be major stakeholders for future ITS deployments which will involve infrastructure support. Road authorities may have the primary responsibility for overseeing the deployment and operation of the following:

• systems that rely on roadway sensors for traffic or weather monitoring; • roadside transponders that enable vehicle-roadside and roadside-vehicle communications; • potentially other infrastructure for co-operative highway-vehicle-control systems.

It is therefore a recommendation of this report that road administrators and engineers obtain a more thorough understanding, through training and other means of information dissemination, of improved warning technology and the status of the research and development underway for other ITS safety technologies so that they are better positioned for future ITS deployments. In this way, the institutional mechanisms that are essential for the successful delivery of rural road safety programmes will be prepared to take full advantage of ITS as a safety-enhancing tool.

VI.9 Conclusion

This chapter has highlighted a variety of rural road applications of ITS in the areas of speed control, driver/vehicle information systems and co-operative systems. In many cases, the technologies are low cost and available now. This means that they can be applied in the rural environment where the extensive nature of the road network makes cost a predominant criteria in decision making.

Given the major role of speed in rural road accidents, the key low-cost applications for rural roads are speed-control technologies such as speed advisory systems and adaptive cruise control. Other near-term, low-cost measures include systems for driver monitoring, intersection approach warning and guide-lights. In the next three to seven years, other low-cost measures such as smart seat-belts and air bags or vehicle data recorders that can lessen the rural road safety problem will be broadly available. Finally, the report identifies ITS measures that are high cost and/or will not be available for some time. Decisions to apply these measures in rural road situations must be made by road authorities and users on a case- by-case basis.

107 BIBLIOGRAPHY

ETSC (1998), “Draft Paper on Telematics and Intelligent Transport Applications for Road Safety”, ETSC, Brussels.

EUROPEAN COMMISSION (Annual Report 1996-97), ITS Applications for Transport, ARTTIC, Brussels.

EUROPEAN COMMISSION (1997), “ITS Transport Telematics -- Meeting the Challenge Together in Europe”, Brussels.

FHWA (1997), Technology for Rural Transport: “Simple Solutions”, Publication No. FHWA-RD-97- 108, Washington, D.C.

FHWA (1998), Intelligent Transport Systems: Real World Benefits, Publication No. FHWA-JPO-98-018, Washington, D.C.

ITS America (1997), A Comparison of Intelligent Transport Systems Progress Around the World Through 1996, Intelligent Transport Society of America, Washington, D.C.

MINISTRY OF CONSTRUCTION (1996), ITS Handbook in Japan, Highway Industry Development Organization, Tokyo.

TRANSPORT RESEARCH LABORATORY (1996), Review of the Potential Benefits of Road Transport Telematics, TRL Report 220, Berkshire.

SWOV INSTITUTE FOR ROAD SAFETY RESEARCH (1997), The Impact of Driver Monitoring with Vehicle Data Recorders on Accident Occurrence, Publication R97-8, the Netherlands.

108 Chapter VII

TRAUMA MANAGEMENT IN RURAL AREAS

VII.1 Special risks of rural road crashes

As shown in Chapter II, the risk of dying in a road crash is greater on rural than on urban roads. Data from Australia (Henderson, 1995) shows that the risk of dying either instantly or before medical attention can be provided increases in line with the remoteness of the location from an urban centre. This essentially comes about because of the increased crash severity in rural areas as a consequence of typically higher speeds and the additional time before treatment is received.

Australian data on the severity of crashes further reinforces these points by showing that in 57% of rural fatal crashes, occupant death is recorded as occurring instantly while for urban crashes the comparable figure is 44%. Likewise, on the basis of a full-scale survey in Hungary (Ecsedy and Hollo, 1994) it was found that in the case of fatalities, about half of all the victims are taken to a hospital before they die. This figure also indicates that about half the fatalities die on the spot or on the way to the hospital. For the purpose of international comparisons of fatality data, the United Nations adopted a figure of 65% of fatalities who died at the scene of an accident or on the way to the hospital (UN, 1994). These examples clearly confirm the greater severity of rural crashes as well as emphasizing the critical importance of time -- and therefore distance -- in regard to emergency help reaching an accident scene. They also highlight the necessity for appropriate mechanisms to transport severely injured victims to distant hospitals and the requirement for adequate medical equipment and personnel at the hospital.

VII.2 Timeliness of treatment

It is logical to assume that the chances of death from injury in a road crash will increase in relation to the time required for the injured person to receive medical treatment, especially treatment of an adequate calibre. An analysis of crash data for rural roads in Australia reveals that the risk of surviving the crash but then dying before receiving treatment appears to be at least 30% higher on rural than on urban roads. In rural crashes, approximately 16% of occupants who had not been killed instantly died before receiving any medical treatment, whereas the equivalent figure for urban areas is 12%.

Typically there are three clear time periods in which death from trauma, including road trauma, can occur. The first period comes immediately in the seconds and minutes that follow the injury. Death is usually due to disruption of the brain, central nervous system, heart, aorta or other major blood vessels. Approximately 50% of trauma deaths occur in this period. Only a few of those patients can be successfully treated and then only in large urban areas where rapid emergency treatment and transport is available. A 1995 study of 155 fatalities in 24 rural counties in the State of Michigan, United States, concluded that about 12.9% of the fatalities could be determined to be definitely preventable or possibly preventable (Maio et al., 1995) if rapid emergency treatment and transport were available.

109 The second period occurs in the one to two hours after the incident. It is sometimes termed the “golden hour”. Death in these instances results from major head injuries (subdural and extradural haematoma), chest injuries (haemopneumothorax), abdominal injuries (ruptured spleen, lacerated liver), fractured femur and pelvis or multiple injuries associated with major blood loss. This second peak occurs in about 15% of injury-producing road crashes, and accounts for approximately 35% of those who die in motorised countries with advanced trauma services. In countries with less-developed services this proportion will be markedly higher. Survival rates are clearly dependent on early and appropriate medical intervention.

The third death period occurs several days or weeks after the initial injury. Major causes of death include brain death, organ failure and overwhelming sepsis. In many of these cases, early treatment may not have made a significant difference to the ultimate outcome.

VII.2.1 Notification, response and arrival times

Response time is the time that elapses between the receipt by the Emergency Medical Services (EMS) of a call for assistance and the arrival of an ambulance on the accident scene. A further time interval has to be added, which is the notification time -- i.e. the time from the accident occurrence until the call is received by the EMS. The arrival time is the sum of the notification and response times. Arrival time is a critical factor as it has been shown (Bernard-Gely, 1998) that the consequences of a crash can be reduced by 1% for every minute saved in the arrival of first aid.

In Germany, in 1995, the mean response time -- the average for urban and rural areas -- to accidents was 8.6 minutes (Schmickler and Joo, 1997). Response times in the United States in rural areas were 12 minutes on average and have not changed since 1982, according to Brodsky (1989). Notification times in rural areas were 9 minutes on average, giving an overall average arrival time of 21 minutes. However, in 14% of urban and rural crashes combined, the arrival time was 30 minutes or longer. These times are clearly too long in life-threatening situations.

Box VII.1 Resource targeting in Finland

The allocation and targeting of resources for fire and rescue services in Finland is based on risk analysis. The purpose of the risk analysis is to assess such factors as the population, the road traffic volume and the frequency of traffic accidents. On the basis of such data, the municipalities are divided into the following four risk areas:

• Risk Area I: Help must be available within 6 minutes of report; • Risk Area II: Help must be available within 10 minutes of report; • Risk Area III: Help must be available within 20 minutes of report; • Risk Area IV: Help must be available within 30 minutes of report.

The criteria for Risk Areas I and II are only met by main routes serving built-up areas. Most rural roads fall into categories III and IV. This example clearly depicts the trauma response difference between urban and rural areas.

A shortening of arrival times can be achieved in a number of ways. For instance, attempts can be made to shorten notification times by wider use of cellular phones and the use of emergency roadside telephones. Response times can be shortened by improving ambulance services or other schemes to improve response logistics, especially in regard to identifying the location of accidents. In this regard, some applications of Intelligent Transport Systems may provide a solution in the not-too-distant future.

110 VII.3 Road trauma treatment in rural environments

As with most life-threatening medical emergencies, the initial assessment and management of a trauma situation will have a significant influence on its outcome because, in a large number of cases, survival depends on the calibre of the initial care. Familiarity with the entire field of medicine is an excellent preparation for the management of the trauma patient. Specific knowledge of basic treatment principles for specific injury types such as head, chest, abdominal or spinal injuries can significantly reduce the morbidity and mortality of trauma patients. Correct, expeditious and efficient management is essential for survival of the critically injured and to prevent aggravation of injuries already received. Properly trained ambulance personnel and adequately equipped vehicles form an essential component in initial management at the injury site and in transit to the most appropriate trauma service. Standing orders and protocols for trauma patients that allow properly trained and certified personnel to initiate life-saving procedures are a necessity in the absence of a doctor. This is especially relevant to road crashes as they provide the majority of cases for emergency trauma management.

Trauma care for people in rural environments poses great difficulties that result from long distances, difficult access and limited financial and professional resources. Historically, poor systems of communication have compounded these problems.

Box VII.2 Improving trauma response and care

In Germany, in 1995, some 195 000 traffic accidents were attended by an emergency team which included a physician. This was about 49% of the total number of accidents attended by an emergency team (Schmikler and Joo, 1997). As well, the response by helicopter to serious crashes in Germany is relatively well developed. In 1995, helicopters started 54 000 times for rescue missions, each operating within a 50 km radius. This was still, however, less than 2% of all emergency missions (Schmikler and Joo, 1997).

Much major road trauma occurs in rural areas, yet the amount seen in each individual rural hospital is small. The first medical officer to see a seriously injured patient in a rural location is likely to be a general practitioner. The more settled rural regions are usually served by a network of community and district hospitals that refer more serious injuries to a major district hospital or base hospital.

Rural trauma services generally serve local communities in country areas. They will not normally have the resources associated with the facilities available in urban and regional trauma services. However, they will be able to provide prompt assessment, resuscitation, emergency surgery and stabilisation, while co-ordinating with the responsible regional trauma service to transfer a patient to an appropriate facility. They require the 24-hour availability of an on-duty medical practitioner as well as a nurse experienced in trauma care.

In the more isolated rural regions, some small hospitals or clinics from time to time receive severely injured patients. They will be unlikely to have pathology services, will have minimal radiology, no intensive-care facilities and no immediately available medical practitioner. The prime need in such services is for early identification of patients needing urgent attention, consultation with the regional trauma service and early transfer of all patients with major trauma to an appropriate centre.

Generally, rural trauma treatment systems are already organised by small groups of dedicated people or by individual general practitioners. They have access to a limited infrastructure. However, these people can organise an efficient trauma-care system and provide effective service when provided with training in advanced trauma skills and relatively inexpensive upgrades in their infrastructure.

111 Optimal care in rural areas can be provided by skilful use of existing professional and institutional resources, supplemented by guidelines that lead to better education and resource allocation. There must necessarily be close linkages between services that provide care in rural areas and those that can offer definitive care for the seriously injured.

Box VII.3 Training medical staff in remote areas

An example from Australia illustrates how the medical staff at hospitals in remote locations can be better equipped with skills to care for victims of road trauma, and serves as a model for other countries. The Australian Advisory Council on Road Trauma in association with the Federal Office of Road Safety organised a workshop at Alice Springs in 1997. Specialist medical practitioners from major urban hospitals presented lectures and demonstrations, with the material being recorded and made available on videotape to a wider rural network of hospitals. The practical nature of the workshop provided a valuable mechanism to address the issue of training of medical personnel in rural areas.

VII.4 Opportunities for improved trauma management in rural areas

While much has been done worldwide to improve the capacity of rural trauma services to successfully respond to the management of road trauma, there does exist considerable scope for improvement. A number of opportunities are discussed below.

VII.4.1 Transport and communications

Mayday systems

Location equipment can serve as the basis for overcoming one of the most vexing problems in trauma management on rural roads, namely the rapid identification of a crash location. When coupled with location equipment through so called “mayday” systems, the actual location of the crash scene can be made available to emergency personnel long before the accident is reported by observers. Mayday systems also have applications for non-crash assistance to disabled vehicles, for instance. The integration of smart technologies into vehicle location systems can also allow emergency personnel to know the severity of the crash before arriving on the accident scene, which, in turn, can better prepare them to provide appropriate assistance. This technology is therefore particularly well suited to help mitigate the effects of lost time in the treatment and transport of crash victims on rural roads (ITS America, 1997).

Early demonstrations of these applications in Europe targeted special vehicles such as taxis and the carriers of hazardous materials (European Commission, 1997). In other uses available on several new automobile models in the United States, the technology is coupled with a cellular phone for automatic dialling. Shibata (1998) reports on several mayday system operational tests and commercial products that are currently available in the United States, as well as related developments in Europe and Japan.

The transport-related elements of arrival time clearly play a major role in determining the outcome of a rural road crash. Thus, any means available to reduce arrival time can improve the safety situation of rural roads. For instance, the lack of information on the exact location of an accident can seriously delay response. Most accidents are initially reported by a passer-by whose knowledge of the road system is limited. One low-cost measure which could be adopted is a more detailed road and kilometre/mile identification scheme that is easily identifiable and understandable to the road user. On busy roads and motorways, some countries go to the length of clearly marking every 100 metre section.

112 More widespread use of Global Positioning Systems (GPS) would help, but it is doubtful that the average motorist will have this in his vehicle in the foreseeable future. A further possible development is automatic crash recorders in combination with GPS. These devices, which have been developed but are not in widespread use, would automatically transmit the exact location once a crash has occurred. These technologies, discussed in Chapter VI, could considerably shorten notification and response times.

In addition to transport, the introduction of improved communications systems for ambulance services, hospitals, medical centres and medical practitioners has the potential to significantly improve the road safety situation in rural areas. As previously mentioned, more widespread use of cellular phones is a relatively simple improvement that can immediately make a difference. One possibility for improving accident-location identification is for emergency services to electronically trace emergency calls made from cellular phones. More advanced communication and technology concepts such as teleradiology should be of high interest to rural areas that want to be prepared to treat road trauma victims. Similarly, direct access telephones should always be available in emergency and operating rooms to permit consultation between doctors in remote and rural areas and specialists at regional or major urban hospitals.

As with speed enforcement, there is a role for publicity in improving trauma reporting and response. Campaigns could be launched that target the general public and inform them of the appropriate steps they can take in the immediate aftermath of an accident that they have been involved in or have witnessed. An important feature of these campaigns would be to educate the public on the type of information that is important and how to communicate it to the EMS. Through such campaigns, the public could, for example, be taught how to clearly and precisely identify the crash location or how to accurately describe the severity of the accident and/or injuries so that an appropriate response can be mustered.

VII.4.2 Standard procedures

Many rural hospitals and clinics face tremendous difficulties when confronted with a serious road-trauma case. These difficulties make treatment far more complex than if the patient were in a higher service urban area. One means to address these challenges would be to establish a set of common guidelines for the care of trauma victims and procedures for case management that could be applied in all rural hospitals and medical centres. The guidelines would serve to assist rural physicians or medical technicians who find themselves in an emergency situation that surpasses the level of their training or experience. Likewise, case-management procedures could save essential time in a critical trauma situation by providing decision support to the medical service professionals who must determine how to handle a given situation in a compressed timeframe. In either case, proper, specific and timely advice could make a difference in saving the life of a trauma victim.

Similarly, the standardisation of equipment and protocols between rural hospitals and the base or regional hospitals to which patients are referred could help to overcome some of the challenges to road-trauma treatment in rural areas. This type of standardisation would help by making the transition from a rural to a regional hospital “seamless” and could, perhaps, contribute to a lessening of rural road crash fatalities.

VII.4.3 Trauma systems

Trauma systems should provide mechanisms that ensure the rapid mobilisation of appropriate clinical support to rural and remote hospitals when the severity of injury exceeds the capacity for

113 treatment locally. The emphasis of trauma services should be on networks and linkages rather than stand-alone trauma centres. Designation of rural hospitals for trauma services will require frequent review, because changes in personnel in regions where staff are generally in short supply will have a great effect on the operation of a given trauma service.

VII.4.4 Training

One means of improving the survivability of rural road crash victims is to provide organised training for all rural doctors and other medical and paramedical personnel in the early management of severe trauma. This type of training can help in the emergency room or at the scene of an accident. In fact, small schemes have been conducted in various countries that set up the logistics so that trained doctors can go to the accident scene. This, however, can be a costly practice unless there is proper screening to ensure that doctors only attend those accidents in which their services are essential.

In order to increase the number of road users able to provide preliminary assistance in a life-threatening situation, there should be increased availability of organised first-aid and resuscitation training for the general population in rural locations. The issue of correct treatment in life-threatening situations is, however, not an easy one for the layman. On the one hand, if there are spinal injuries, the patient should be moved only carefully and expertly. On the other hand, if lack of breath or massive bleeding are the case, immediate action can be taken and could prove to be lifesaving. In rural areas these issues deserve more detailed attention. Of course, any training of this type must include information related to liability issues and special personal risks to be avoided by amateur rescuers trying to assist crash victims.

VII.5 Conclusion

The risk of dying in a rural road accident is considerably higher than for a similar accident in an urban area. There are several causes for this, but they all revolve around three main issues: the timeliness of response; the adequacy of initial care; and the ability of rural hospitals to administer proper care to severe trauma cases. There are three timeframes when road-trauma fatalities occur. In the first time period, death is immediate or in transport. In these cases, arrival time and immediate care, including rapid transport to a hospital, are essential. As many as 35% of the deaths on rural roads occur in the second time period which is known as the “Golden Hour”. In this case, rapid and reliable response coupled with sufficient care at a rural or regional hospital is essential to saving the life of the victim. Several suggestions were provided for improving survivability in both of these timeframes.

Identifying an accident location is one of the key problems in responding to rural road crashes. Several options are recommended that can improve the situation, including: upgrading road and kilometre/mile identification schemes; expanding the use of GPS; and exploring possibilities for automated accident detection. Several communications technologies should also contribute to improving rural road safety. Paramount among the available technologies is cellular telephones which can shorten arrival time and improve the overall information available about an accident situation. Some consideration should be given to how publicity campaigns can be used to better inform the public about how best to react at the scene of an accident -- i.e. how to communicate to the EMS, what kind of information to provide and first aid that can be administered immediately.

Common guidelines and standard procedures at local hospitals can aid in preserving the lives of trauma victims. Finally, training at all levels, from the medical practitioner to the general public, could play a far more important role in saving the lives of trauma victims in rural settings than in urban areas.

114 BIBLIOGRAPHY

BERNARD-GELY, A. (1998), Building on a Data Foundation: France’s National ITS Policy in Focus, Traffic Technology International, June/July, UK and International Press, Surrey.

BRODSKY H. (1989), “Evaluating Emergency Medical Service Arrival Time in Road Accidents, USA”, in Proceedings of the First International Conference on New Ways and Means for Improved Road Safety, Transportation Research Institute, Technion, Haifa, Israel.

ECSEDY G. and HOLLO P. (1994), “The first Hungarian national traffic safety program and experience of its implementation”, in Proceedings of the third international conference on new ways and means for improved road safety, Transportation Research Institute, Technion, Haifa, Israel.

EUROPEAN COMMISSION (1997), “ITS Transport Telematics -- Meeting the Challenge Together in Europe”, Brussels.

HENDERSON, M.A. (1995), Report of the Wodonga Seminar, Rural Road Safety: Focus for the Future, National Road Trauma Advisory Council, August.

ITS America (1997), A Comparison of Intelligent Transport Systems Progress Around the World Through 1996, Intelligent Transport Society of America, Washington, D.C.

MAIO R.F.,GREGOR M.A. and WELCH K.B. (1995), “Preventable Trauma Deaths in Rural Michigan”, 39th Annual Proceedings of the Association for the Advancement of Automotive Medicine (AAAM), Chicago.

NATIONAL ROAD TRAUMA COMMITTEE (1989), Early Management of Severe Trauma (Course Manual), Royal Australasian College of Surgeons, Canberra.

NATIONAL ROAD TRAUMA ADVISORY COMMITTEE (1993), Report of the Working Party on Trauma Systems, Australian Government Publishing Service, Canberra.

SOUTH AUSTRALIAN HEALTH COMMISSION (1988), Report of the Trauma Services Review Committee, Adelaide.

SCHMIKLER, M.R. and JOO S. (1997), “Emergency Medical Services in Germany: Data on 20 Years of Development”, in Proceedings of the Third International Conference on New Ways and Means for Improved Road Safety, Transportation Research Institute, Technion, Haifa, Israel.

SHIBATA, J. (1998), General Trends of ITS in the World: Key Technology to Improve Traffic Safety, IATSS Research, Volume 22, No. 2.

UN (1994), “Statistics of Road Traffic Accidents in Europe”, UN Economic Commission for Europe, Geneva.

115

Chapter VIII

STRATEGIC FRAMEWORK

VIII.1 Introduction

Road safety problems outside urban areas have unique characteristic features as well as possibilities and limitations for solutions such that it is often difficult to adapt motorway or urban-based approaches. For instance, rural roads and road networks have developed over the years and have been adapted to changing conditions, growing mobility and motorisation in a step-by-step manner. In other words, the rural network has evolved rather than being designed. This important feature clearly separates the rural network from motorway networks. Because of this feature, rural roads have many different functions and inconsistencies. The essential first step for addressing road safety on these roads is, therefore, to determine how the different functions of these roads can be safely combined, especially given the relative high speeds that prevail outside urban areas. Section VIII.2 therefore presents a vision, from a road-safety perspective, of how these functions can be combined.

The chapter also explores decision making in relation to realising a safe road network outside urban areas. Decision making in this regard is critical because it is not sufficient to individually consider the different knowledge building-blocks presented in Chapters IV through VII. Rather, in reaching a decision, the possible measures could, and perhaps should, be weighed against one another. Furthermore, not only safety concerns are at stake when decisions are made. Therefore, safety should be considered in a broader context that incorporates the social and economic consequences of a given decision. Thus, while the approach within an urban area is based on a general concept, the improvement of rural road safety cannot usually be handled in the same manner. In Section VIII.3 possibilities are elaborated for a more integrated and rational decision-making process for safety improvements on the rural road network. Institutional consequences are discussed in Section VIII.4. Finally, conclusions are presented in Section VIII.5.

VIII.2 The rural road network: a functional approach

Roads are built with one major function in mind, that is to enable the movement of people and goods -- the so-called travel function. Beyond this, three options can be distinguished:

• the flow function on through roads: enables high speeds for long-distance traffic and, often, high volumes; • the distributor function: serves districts and regions containing scattered destinations; • the access function: enables direct access to properties alongside a road or street.

As discussed in Chapter III, a combination of functions on a single roadway leads to higher accident rates. This can generally be understood by considering the effects caused by having one road used by road users with different intentions and, therefore, with different kinds of behaviour. Additionally, road authorities anticipate these different behaviours by taking the differences into account

117 in the road design. This situation leads to road designs that initiate with divergences -- i.e. concepts for roads with multiple functions -- which ultimately make the road user uncertain as to the intention of certain roadway features. This also explains the relatively high accident risks on rural roads in comparison to roads and streets where the idea of mono-functionality with a corresponding design serves as the starting-point.

On the basis of a functional assignment, it is possible to structure a categorised road network. This is not something new and is fairly well known, at least in theory. The present practice certainly includes the belief that a certain combination of the flow, distributor and access functions takes place on almost every road and street. However, from the viewpoint of safety, it is better to structure a categorised road network on the basis of mono-functionality. The differences between the present practice and the practice that is more inclined to road safety -- i.e. sustainably safe practice -- are shown in Table VIII.1.

Table VIII.1 Common and sustainably safe practices for categorising roads and streets

Common practice today Sustainably safe practice

Existing types of Sustainably safe roads Traffic function types of roads

Motorway Ia. Motorway Increasing Flow Motor road • flow and Ib. Motor road decreasing or Main distributor Access IIa. Distributor road (rural) Local distributor distributor IIb. Distributor road (semi-urban) District artery IIIa. Access road or (rural) Neighbourhood artery Decreasing Residential street flow and access IIIb. Access road • increasing (urban) Residential area access

Residential function

Source: Wegman and Elsenaar, 1997.

If absolutely dividing these functions is practically impossible or requires an excessively high investment, then it is recommended to place the priority at the “lowest” function. For example, if a rural road should combine a flow function and a distributor function, then the design should have as its starting-point the requirements of a distributor function. An example of a categorised road network is shown in Figure VIII.1.

118 Figure VIII.1 Categorised rural road network

Westland - Sustainably safety scenario

Through Distributor Access

meters

0 1000 2000 3000 4000

VIII.3 Rationalising policy making

In the areas of traffic, transport and road safety there is an increasing interest in rationalising policy making. This means that, in economic terms, scarce resources ought to be allocated in a rational way in order to maximise societal welfare. This implies that all possible effects of certain decisions must be recognised, along with their costs. A certain value should then be given to the benefits and costs, thus allowing the different options to be weighed against one another. As shown in Box VIII.1, one means available to accomplish this is the application of safety-management systems.

For decisions concerning investments to improve rural road safety, this theoretical approach is not always followed and, in fact, sometimes cannot be followed. When considering investments, not all costs and benefits are taken into account, on the one hand, because they are not all known and, on the other hand, because they do not all have to be carried by the investor. For instance, the costs of traffic accidents are seldom, if ever, carried by the road authority just as prevented accidents form no benefit even though the investments for accident prevention are financed by the road authority. Likewise, those who profit from fewer accidents -- i.e. road users and, for example, automobile insurers -- generally do not bear the investment costs for reducing the chances for or seriousness of accidents. However, benefit accrues to society through the lower medical and hospitalisation costs.

119 Box VIII.1 Safety management

As effective as specific highway infrastructure development and operational improvements can be in improving road safety, they cannot address all road safety problems. A systematic approach to highway safety problem resolution should be taken to provide direction for safety programmes to address problems identified for each component of the transportation system: the human (drivers, occupants, pedestrians); the vehicle and the roadway. The term “safety management” has been used to describe an approach for safety that is similar to established programmes such as Pavement Management Systems, Bridge Management Systems and Information Management Systems. It is important to note that safety management differs from other management systems because it relies first and foremost on the creation of a team of all those who have a stake in road safety -- e.g. safety engineers, police, emergency medical technicians, hospitals, etc.

A description of safety management can be found in the United States Federal Register (1996). The concept was further elucidated by the United States Federal Highway Administration (FHWA, 1997). The United States General Accounting Office provided related information on safety management in a report (GAO, 1997) that described how individual states were implementing various management systems. Case studies were also developed by the FHWA (1998) to document successful safety-management practices in several states. ITE (1993) provides a description of safety-management systems, their component and development.

While the underlying safety management concept is not new, many safety programmes have been developed independently of each other, rather than systematically, since safety functions can reside in a number of different national, state or local agencies. As mentioned earlier, the focus must be on the effective management of the system, and the driver, vehicle and road inter-relationships as a whole. By introducing the proper framework, safety management can provide a mechanism for more efficient use of limited resources as well as ensuring that various highway safety disciplines work together to achieve common goals.

Safety management objectives could include:

• establishment of long- and short-term goals; • establishment of accountability; • recognition of institutional and organisational initiatives; • collection, maintenance and dissemination of data; • analysis of available data; • evaluation of the effectiveness of safety activities; • development and implementation of public information and education activities; • identification of skills and resources to implement the programmes.

The recent Highway Safety Strategic Plan (AASHTO, 1998) in the United Sates is a good example of establishing systematic (driver-vehicle-road) strategic goals as part of safety management at the national level.

VIII.3.1 Benefit-cost analysis

The profitablility -- i.e. financially for businesses or economically for society -- of road safety investments and accident prevention has been clearly identified. Studies by government-owned automobile insurers have indicated that several road safety measures can be financially profitable (Johnson, et al., 1997). Some examples where these companies exist and such studies have been performed include: the state of Victoria, Australia, with the Transport Accident Commission (TAC); the province of Québec, Canada, with the Société de l’Assurance Automobile de Québec; and the province of British Columbia, Canada, with the Insurance Corporation of British Columbia. The most recent version of the Norwegian Road Safety Handbook (Elvik et al., 1997) also includes many examples of safety

120 measures with a benefit-cost ratio higher than one, which means that the measure is profitable to society. Similarly, the BTCE (1995) conducted an analysis of the Black Spot Programme in Australia and concluded that such a road safety measure had a benefit-cost ratio greater than four. Finally, in an early draft of a Dutch Road Safety Plan (Elvik, 1997), benefit-cost ratios were developed for several measures which were under discussion at that time. The resulting figures provided general indications as to the profitability of the different measures.

Although it is unrealistic to justify investment decisions on economic criteria alone, this type of information remains important as a basis for informed decisions. Benefit-cost information is primarily important in the context of rationalisation. For instance, if one is restricted in policy making to only road safety, then a work method can be envisaged where a “socially and politically” acceptable goal is set -- for example, a 25% reduction in the number of traffic casualties within ten years. It would then follow that only those measures are selected that will both help to achieve the goal and result in the lowest costs to society. This would be a cost-effectiveness approach. Naturally, one should take into account developments that can be expected if no additional policies are established, if there are certain restrictions that cannot be influenced, if there are undesired side-effects from regulations or other such concerns (OECD, 1994). An example where the cost-effectiveness approach is followed is a recent road safety plan in Finland (Ministry of Transport and Communication, 1997). The Finnish plan estimates that for the period 1997-2000, FIM 3.3 billion (USD 609 million) is needed, of which FIM 2.1 billion (USD 388 million) should be raised by public bodies and FIM 1.2 billion (USD 221 million) by Finnish road users. The programme’s benefit-cost ratio will be 3.1 when its entire period of effect is taken into account.

Another example to illustrate the development of rational road safety policy is the recent policy paper from the European Commission (EC) on road safety (European Commission, 1997) in which there is a presentation of what has come to be called the “One million euros Test”. The EC suggests that when making decisions on proposed measures, the expected costs should be compared with the economic benefits that result from accidents that are prevented due to the measures. A rough estimate shows that on economic grounds alone, one million euros can be invested in measures if one fatal accident -- and the corresponding number of injury and damage-only accidents -- can be prevented. The figure of one million euros should not be taken too literally and applied in all OECD Member countries because some theories suggest that this is an underestimation -- e.g. the valuation of the lost quality of life was excluded, under-reporting of accidents was not taken into account and severity levels other than fatalities were disregarded. Caution is also called for because there are different costs -- i.e. for accidents or valuation of a lack of road safety -- among countries for providing an improved level of safety. For example, the BTCE (1996) recently reviewed different approaches to measuring transport safety, including the valuation of loss of life and property, to provide a framework for assessing policy options for safety.

The most recent developments in road safety policy in Australia are based on benchmarking of initiatives, penalties and enforcement regimes to determine which policies have achieved the greatest gains in reducing fatalities and serious injuries. Policies are being compared across states and territories with the aim of lifting performance in those areas which have lagged in realising relative gains by adopting those strategies which have yielded significant results.

VIII.3.2 Integrated safety policies

Decision making is, of course, more complicated if other aspects in the policy-making process are included. In a recent research report (OECD, 1997b), an overview is given of possible strategies for

121 the integrated consideration of road safety and environmental policies. This type of consideration is important if the action to improve one area creates a risk that the other will be adversely affected. However, the problems of integration do not lie solely in the expansion of the sphere of action, they also encompass the increasing number of criteria and actors involved. In the report, some key factors are identified which possibly contribute to the success of integration: “an appropriate choice of public involvement, a detection of issues with special political or other connotations, an understanding of the paradoxes in what people will accept, a fair consideration for points of no return, a definition of possible packages of transport system measures, an identification of the gainers and losers from particular decisions and the possibility to learn from failures as well as successes”. It is recommended to use a common set of indicators and to carry out impact evaluation studies “in the other sector”. The report also recommends that safety and environmental balance-sheets should be considered as a normal part of road transport planning. This requires that in the field of road safety, a road safety impact assessment or similar analysis tool be developed.

Unfortunately, the limitation to only safety and environmental issues means that other relevant aspects are not considered. Elvik (1997) gives a more integral overview of the most important aspects and their economic costs (Table VIII.2).

Table VIII.2 Relevant effects in benefit-cost analyses of transport policy and their economic valuation

Major policy objective Aspects of policy objective Relevant economic valuation

MOBILITY Travel time Costs of travel time Travel time regularity Costs of traffic congestion Transport costs Vehicle operating costs ENVIRONMENT Air pollution Costs of air pollution Dirt and dust pollution Costs of dirt and dust Traffic noise Costs of traffic noise SAFETY Number of accidents Costs of accidents ROAD STANDARD Engineering standards Road investment costs Maintenance standards Road maintenance costs

Source: Elvik, 1997.

Within this context, benefit-cost analyses should be carried out in order to support decision making that leads to the “maximum welfare” for society. In this scenario, efficiency can be measured with the so-called Pareto criterion which stipulates that a project improves welfare if it makes at least one person better off and nobody worse off. By using a balance sheet, this kind of analysis could be undertaken.

A recent study (OECD, 1997a) defines several relevant indicators that can be taken into account when deciding on investments in the road-traffic sector. The indicators go beyond accessibility/mobility, traffic safety and environment to include government and road administration indicators for such things as programme development, programme delivery and programme performance. The underlying principle is that public administrations operate in a complex political environment and it is no longer valid to simply deliver services to society. The views of road users and the influences of peer groups are considered to be far more relevant than they were in the past. Furthermore, road investments have to contribute to societal

122 goals for equity, community and economic development. A conceptual model for performance indicators was developed and is being tested in field trials in 15 OECD countries to ensure that the end-product of this study can be used in day-to-day decision making.

It becomes clear in reviewing the approaches described above that one cannot assume that political decisions can simply be made based on the results of calculations. These decisions also require social involvement and appreciation. In respect to safety problems on rural roads, it is equally important to consider the different branches of society that have a stake in the outcomes of decisions. In the first place, there are the different categories of rural road users -- commercial, agricultural, private, etc. -- with almost every category having its own lobby. One must also consider the private sector, which in this context includes the insurance industry, the transportation industry, the industries responsible for construction and maintenance of roads, and others. Finally, there is the government in its many forms. Naturally, a question could be raised in relation to the optimal nature of the decisions that are made within the government. Concerning road safety, the following government elements are involved:

• that section of the government pertaining to general road safety problems, the international aspects of policy making (vehicles, many ITS developments), regulations (highway code, licensing), involvement with other authorities (e.g. subsidies), and other relevant matters; • that section of the government which concerns itself with influencing the behaviour of road users (police, judiciary); • that section of government dealing with trauma management and health services (emergency medical services, hospitals).

The approach presented here is of course formalistic and of an institutional nature. Obviously, the formal responsibility of organisations and agreements between organisations gives an important framework for effective and efficient policy making. But it would be wrong to think that this formal structure explains everything. The role of personal relationships and informal contacts is more important than is generally described in textbooks and research reports.

The desire to rationalise policy making in the field of road safety and to make decisions in a more integrated manner is not very easy to fulfil. It is recommended to choose a logical line in decision making as a starting point for policies, although in reality it must be admitted that a logical line may not always be reflected in the policies that are implemented. It is further recommended to always be very accurate in defining frameworks within which the policies are enacted. Given all of the possible intended or unintended consequences, calculations could be theoretically correct but in practice lead to indecisiveness and infeasible research.

What does more rational policy making mean in terms of the chances for and threats to decisions to improve safety on roads outside the urban areas? First of all, we are dealing in this instance with problems which have a relatively low political priority when compared with problems on main roads (congestion), the revitalisation of cities, the attention given to transport-related problems of the environment, the attention for public transport, and other pressing issues. This means in all probability that the improvement of safety on rural roads is more a question of step-by-step progress -- hopefully many small steps -- rather than large-scale concentrated improvements. The second conclusion could be that integrated policy making, given the fact that many actors are involved or in this case should be involved, is not easily attained and requires communication, co-ordination and co-operation (see Section VIII.4).

123 Measures in the areas of infrastructure and ITS applications are often motivated by problems other than safety problems that must be resolved. ITS applications often target the ever-important quality improvement of (economic) accessibility. Infrastructural measures address the accessibility issue, but go beyond this to adapt the road infrastructure to new circumstances -- i.e. newly built locations, new infrastructure, etc. -- or to meet “regular” road maintenance requirements. If, when considering different measures, it is decided, for whatever reason and in whatever manner, to influence the traffic flow, then it should be required to explicitly consider the road safety consequences. Should the assessment of the safety consequences not be satisfactory, then the decisions have to be adjusted so that a safer result can be obtained. This requirement does not pertain only to infrastructure decisions, but also to decisions for influencing behaviour, for deploying ITS applications and for organising trauma management.

In the area of implementation of the concept of explicitly considering road safety concerns in decision making on the structure and the design of the road networks, there are a few interesting examples to be given (Wegman et al., 1994; ITE, 1994; and ETSC, 1997). The ETSC report describes two procedures -- Road Safety Audit and Road Safety Impact Assessment -- as follows: “A road safety audit is a formal procedure for independent assessment of the accident potential and likely safety performance of a specific design for a road or traffic scheme, whether it is new construction or an alteration to an existing road. A road safety impact assessment is a formal procedure for independent assessment of the likely safety effects of proposed road or traffic schemes that have substantial effects on road traffic, upon the occurrence of accidents throughout the road network upon which traffic conditions may be affected by the schemes.” The two procedures are complementary in that the aims are similar, but there are differences in their scope and timing. The basis for a safety audit is the application of safety principles to the design of a new or a modified road section. On the other hand, a road safety impact assessment should be carried over a larger part of the road network to examine the impact of larger schemes.

For the other policy areas such as enforcement and trauma management, there are no known examples as to which way the decision making on priorities could be influenced so that road safety is included in a satisfactory manner. It is therefore recommended that audit or impact assessment procedures be developed for these policy areas.

The approach to road safety with regard to policy making consists of a standard series of steps (stages or stadia) that are generally followed in most parts of the world. If the original approach to improve road safety in a location consisted of tackling segments of problems in a rather isolated way, it was generally realised after a period of time that the road traffic system would inevitably contribute to a large number of accidents and casualties. Under these conditions, the system itself is intrinsically unsafe. When using an analysis of road safety problems with an eye to considering the potential of individual measures, it is necessary to determine which policy phase a country or other jurisdiction has achieved. In general, there is currently a movement towards a more integrated approach in the planning phase, at least as far as the organisation of policy implementation is concerned, particularly if there are different authorities involved. Integration makes it easier to reach the goals with lower costs and without detrimental side-effects.

VIII.3.3 Integrated rural road safety policy -- possibilities and options

In Section VIII.2 it was observed that roads outside urban areas have to fulfil three different functions: a flow function, a distributor function and an access function. That section also makes a plea to strive as consistently as possible for granting only one function -- i.e. mono-functionality -- to each road or street. Then, for each functional category of road, different possibilities exist for improving road safety. Table VIII.3 presents a matrix that could be used to chart the possibilities for improvements of the

124 various road categories. It is of importance to be aware of the necessity to approach every cell in the matrix respectively and to acquire validated knowledge on effects and costs of all potential measures. This would then be a starting-point for developing packages of measures and for integration.

This type of approach will only be successful if it builds on a road safety vision and a strategy derived from this vision. (See, for example, the recent visions developed in the Netherlands on sustainable safety and in Sweden on zero-vision.) Such visions and strategies have to be adapted to regional circumstances. Based on a vision and a strategy, an integrated road safety policy could be developed. It is recommended that these visions and strategies be developed on a nation-wide basis and, as a result, that each sector commits itself to certain improvements. An example of such an approach is given in Australia (National Road Trauma Advisory Council, 1995). Another example is the Strategic Highway Safety Plan in the United States (AASHTO, 1997).

Table VIII.3 Matrix to be used in an integrated approach for improving road safety

Functional Road infrastructure Legislation and ITS/Telematics Trauma management category enforcement Flow road Distributor road Access road Source: OECD.

In Chapters IV, V, VI and VII, an overview is given of the different possibilities for each cell. How does one then make a choice between the cells? First, the potential contributions in each cell have to support the vision and strategy. Secondly, estimations should be made of the cost-effectiveness per measure. Based on these two steps, packages of measures could be composed in which possible contributions are considered from all of the cells. The time perspective of the measures -- i.e. When could a measure be effective? How long will an effect last? Will an effect erode over time? etc. -- is of great importance and should be considered. In addition, when it comes to implementation of certain measures, it is not always easy to see the added value of overall integration, though every effort should be made in this direction.

It is advisable that two more dimensions be added to this matrix. First, the accident, types, as defined in Chapter II, could be used to define a number of “target accidents” that could be changed by a certain countermeasure and could then be added. Second, it is possible to add a time dimension to the matrix. For example, measures in the field of enforcement could possibly have their positive safety effect immediately, while gains from ITS measures may only appear after ten years.

Trauma management as a sector to improve safety on rural roads should not necessarily be integrated in the other sectors. It has to be considered as an important field to reduce the health consequences and economic costs of road accidents, but it can be considered as a stand-alone measure. The effects are more or less independent from other effects as it is not possible to replace trauma management measures by any activity in one of the other cells. The effects of trauma-management should also be measured in terms of influencing societal costs of accidents rather than in terms of changing the number of accidents or casualties.

ITS applications developed for rural roads have to answer road safety problems specific to these roads. Better problem statements and analyses are required to direct research and development in this field. However, ITS research and development takes place in another (more international) arena and it is not always possible to integrate these developments with road infrastructure and enforcement in a

125 particular region. Of course, certain parts of a rural road network could be used as a test-site when developing certain telematics applications.

Infrastructure and police enforcement (including public information) are absolutely essential in policy integration. In this case, it is recommended to organise integration on a regional basis in which all road authorities, police and judiciary are co-operating with partners from the private sector. On flow roads, and in some instances on distributor roads, this co-operation can help to detect locations where there are high crash risks. Beyond this, there are local activities to be developed in the area of road design and, likewise, with regard to traffic control. This approach is less appropriate on access roads and on some distributor roads simply because high-risk locations cannot be detected. There are, of course, exceptions in a few locations where obvious design mistakes or local unsafe behaviour can be identified. In general, however, on these flow roads, measures can be thought out in an efficient manner. On the other two road types -- i.e. flow and distributor -- a local approach is not always possible and, as is the case with problems in urban areas, a network approach must be taken.

VIII.4 Policy: organisation -- financing -- information/data

The key to a more integrated approach for addressing the road safety problems on rural roads lies in the idea of creating “shared interests”. In order to truly have shared interests, a sound definition and interpretation of the notions of communication, co-ordination and co-operation between different government organisations, the private sector and road-user organisations must be in place. Willingness to co-operate is of utmost importance.

Box VIII.2 Rural road safety action plan: An example from Australia

Based on evidence that gains in road safety in rural areas were not keeping pace with those in urban areas and the disturbing fact that country drivers were twice as likely to be involved in a fatal crash as city drivers, the Australian Transport Council agreed to support the development of a Rural Road Safety Action Plan in 1996. The plan was developed by 22 member organisations representing the federal, provincial and local governments, police organisations, surgeon and health-service organisations, road-vehicle manufacturers, and engineering institutes. The plan resulted in 24 specific action recommendations in the following activity areas:

• planning road improvements; • public education programmes; • involvement of local communities; • speed management; • fatigue management; • enforcement; • trauma services; and • remote areas. A premise for the plan is that all member organisations share responsibility for implementation. Thus, a lead agency and supporting agencies were identified for each specific action recommendation.

Aside from the managerial approach outlined above, there are two important conditions which have to be fulfilled for the succes of an integrated safety policy. First, participants involved in the (co-ordination of the) implementation of policies should be, for the most part, “road safety professionals”. Secondly, it is recommended that all partners should have relevant and reliable information at their

126 disposal particularly in monitoring implementation of initiatives and the gains resulting from those initiatives. An example of such an approach is given in Box VIII.3.

Box VIII.3 Road safety information system (RIS): Key information supporting traffic safety policy in the Netherlands

The Dutch road safety information system is part of a policy-making concept: the concept of rational decision making. The main benefit of this system is the user-friendly access to actual, relevant and qualified information, which can also easily be processed in e.g. documents and spreadsheets. The selection of information in the system is agreed upon by the users.

Information in the system enables quick understanding of the developments by the parties concerned and timely adjustment of policy. The advantages of the RIS are:

• it integrates information from various sources; • it contains recent information and the best there is in the Netherlands; • the user can request every required combination of data; • all data in the RIS are accompanied by explanatory text; • any question that is not answered by the RIS can be put to the RIS Information Desk.

The RIS covers the complete range of spearheads of official national traffic safety policy. For each topic, a time series of data is available. Within the possibilities of each table, the user can choose the variables and the classes to be presented and the level of detail. Of course tables can also be presented as graphs. More than 60 indicators are available.

Starting the Dutch RIS, the user is only two screens away from the information he wants to see. The first screen offers all subjects, one of which is to be selected (Indicator Selection Screen), and the second screen enables menu-driven specification of the requested information (Indicator Manipulation Screen). The output screen contains buttons for further processing of the information.

Concerning the financing of policies and measures, it is reasonable that the implementation of tasks should be financed by the partner who has the responsibility and authorisation in a certain area. For instance, the road authorities finance investments in the infrastructure, the police and judiciary finance the enforcement, the health services (or health insurance) finance the trauma management, and so on. In an integrated approach there should also be agreement regarding the finances, especially where benefits accrue to all partners. An interesting example in this regard involved speed cameras in Queensland, Australia. It was agreed that these cameras could be introduced as long as any revenues generated from their use was dedicated to road safety programmes rather than for resourcing the Queensland police. It is recommended that separate financing be agreed upon for the co-ordination of policies, for the information facilities and for the education of road safety professionals.

A very interesting approach might be to link road safety improvements with maintenance investments and with developments in other areas relevant to road safety. If maintenance budgets are the carrier of safety improvements, the costs of these safety improvements are nearly invisible. It is not necessary to look for “new money” if one would like to invest in order to improve road safety. Especially under conditions of reducing budgets, it is advisable to redirect existing budgets. The re-arrangement of present budgets demands an integrated consideration of targets and thereby suitable measures and countermeasures. Thereafter, measures could be ranked according to their economic benefits in relation to the expected costs. One point of consideration in this regard is that those who currently invest in the improvement of road safety -- mostly the government and therefore the taxpayer -- are not those who profit from the lower costs of road accidents. For the most part, insurance companies -- accident, vehicle,

127 labour, disability, health, and life insurers -- and therefore the premium payer, are the beneficiaries of these lower costs. However, citizens are not only taxpayers but also premium payers. This raises an interesting question for the future, namely: Is it possible to create a “single purse” for the investors in accident and injury prevention and those who benefit from less economic detriment as a result of fewer accidents?

VIII.5 Summary

Accidents on rural roads demand a separate and specific approach because their characteristics and the possibilities for preventing them differ greatly from motorways and roads and streets in urban areas. The following three elements are key ingredients for an effective and efficient approach to improving the safety of rural roads.

1. From the point of view of road safety, it is essential to have a hierarchically structured rural road network based on functional requirements. It is not safe to combine different types of road use and relatively high driving speeds if in fact road design and road behaviour is not attuned to this. This implies that there should be a fundamental reconsideration of the design and use of roads outside urban areas in many countries.

2. When striving for a safer reconstruction of rural roads, it is important to make an integrated evaluation among all the involved interests -- i.e. mobility, environment, integration in present conditions and circumstances, costs, etc. -- and to make the evaluation as rational as possible. The development of rational decision-making models and the attainment of relevant data is therefore recommended. The actors (the organisations) included and the decision-making process chosen should be based on a goal for integrated and rational decision making. Although the quality of the decision making can be better carried out in this way and better decisions will result, it requires a more complex decision-making process than is now the case.

3. A challenge is to determine how organisations which form a part of the decision-making process can reach consensus on more integrated decision making. It is therefore suggested that the senior management of organisations responsible for planning, infrastructure, legislation, enforcement, ITS development and trauma management, in co-operation with political leaders, be requested to initiate this process. It is expected that, if this integrated and rational decision-making approach is achieved, a substantial improvement in rural road safety could be realised.

128 BIBLIOGRAPHY

AASHTO (1997), Highway Safety Strategic Plan: A Comprehensive Plan to Substantially Reduce Vehicle-related Fatalities and Injuries on the Nation’s Highways, Washington, D.C.

BTCE (1995), Evaluation of the Black Spot Program, Report 90, AGPS, Canberra.

BTCE (1996), Valuing Transport Safety in Australia, Working Paper 26, DOTRD, Canberra.

ELVIK, R. (1997), A Framework for Benefit-cost Analysis of the Dutch Road Safety Plan, TØI, Oslo.

ELVIK, R., MYSEN, A.B. and VAA, T. (1997), Trafikksikkerhetshandbok, Transportøkonomisk institut, Oslo.

EUROPEAN COMMISSION (1997), Promoting Road Safety in the EU, The Programme for 1997-2001, Commission of the European Communities, Brussels.

EUROPEAN TRANSPORT SAFETY COUNCIL (1997), Road Safety Audit and Safety Impact Assessment, ETSC, Brussels.

EUROPEAN TRANSPORT SAFETY COUNCIL (1997), Transport Accident Costs and the Value of Safety, ETSC, Brussels.

FEDERAL REGISTER (1996), “Management and Monitoring Systems; Final Rule”, Part III, Department of Transportation, Federal Highway Administration/Federal Transit Administration, Washington, D.C.

FHWA (1997), “Case Studies of Highway Safety Management Systems”, Federal Highway Administration, Washington, D.C.

FHWA (1998), “Safety by Design: Office of Highway Safety Business Plan for 1998-1999”, Federal Highway Administration, Washington, D.C.

GAO (1997), “Transportation Infrastructure”, States Implementation of Transportation Management Systems, General Accounting Office, Washington, D.C.

ITE (1993), The Traffic Safety Toolbox: A Primer on Traffic Safety, Institute of Transportation Engineers, Washington DC.

ITE (1994), “Informational Report: Road Safety Audit”, Committee 4S-7, Institute of Transportation Engineers, Washington D.C.

129 JOHNSON, M., NEPOMUCENO, J., ZEIN, S., and YEE, H. (1997), An Evaluation of ICBC-funded Road Improvement Investments, Proceedings of the Canadian Multidisciplinary Road Safety Conference, Toronto, Ontario.

MINISTRY OF TRANSPORT AND COMMUNICATIONS (1996), “Road Safety Plan, the Recommendation of the Consultative Committee on Road Safety”, Ministry of Transport and Communications, Helsinki.

NATIONAL ROAD TRAUMA ADVISORY COUNCIL (1995), Towards an Action Plan for Rural Road Safety, a report of the Wodonga seminar “Rural Road Safety: Focus for the Future”.

OECD, ROAD TRANSPORT RESEARCH (1994), Targeted Road Safety Programmes, OECD, Paris.

OECD, ROAD TRANSPORT RESEARCH (1997a), Performance Indicators for the Road Sector, OECD, Paris.

OECD, ROAD TRANSPORT RESEARCH (1997b), Integrated Strategies for Safety and Environment, OECD, Paris.

OECD, ROAD TRANSPORT RESEARCH (1997c), Road Safety Principles and Models, OECD, Paris.

WEGMAN F. C. M., ROSZBACH, R., MULDER, J.A.G., SCHOON, C.C. and POPPE, F. (1994), Road Safety Impact Assessment, SWOV, Leidschendam.

WEGMAN, F.C.M. and ELSENAAR, P.M.W. (1997), Sustainable Solutions to Improve Road Safety in the Netherlands, SWOV, Leidschendam.

130 Chapter IX

CONCLUSION AND RECOMMENDATIONS

IX.1 Rural road safety -- a sleeping giant

Rural roads are defined in this report as roads outside urban areas that are not motorways or unpaved roads. Rural roads can be classified by function, by administrative category or using a combination of the two. The report concludes that the wide variety of principles and implementation practices used in road classification schemes hampers and obscures a correct representation of the size and nature of the rural road safety problem and makes it difficult to compare rural road safety across countries.

Rural road safety problems suffer from a lack of attention compared to the safety problems on motorways and urban roads. It is quite evident that this situation is inappropriate given both the extent and relative magnitude of the problem. Lack of attention is found both in governmental organisations and in the general public. Some of this inattention may be explained by the fact that accidents in urban areas create a certain amount of subjective feeling among citizens because a single urban accident is publicly more visible than ten accidents spread over the vast rural network. These urban accidents or near-accidents appeal to the imagination because they are considered rather close and personal to many people. Special attention is also usually granted to major accidents on motorways, especially serious accidents that involve several vehicles. These accidents also cause serious congestion which generates a high degree of media exposure.

Accidents on rural roads do not generally possess the features that catch the attention of the general public or the media. As well, there is widespread belief that preventing rural road accidents is extremely complex because rural road accidents are incidental, far-away events that are caused by unsafe individual behaviour that cannot be influenced by proper road design or effective police enforcement. In other words, “the lonely driver is to blame”. Although this might correctly describe what people think, this report contains enough evidence to refute this view and urges strong new actions -- many summarised in the following sections -- to reduce the rural road safety problem.

IX.1.1 Rural road safety problem

Rural road safety accounts for a considerable share of the total road safety problem. Each year, more than 75 000 people are killed on rural roads in OECD countries. This equals more than 60% of all road fatalities in OECD countries. The economic costs of this safety problem are staggering -- on the order of USD 135 billion (approximately 120 billion euros) per year.

The increasing safety qualities of automobiles have contributed to a certain reduction in the number of road crash fatalities, even in severe crashes. It is likely, therefore, that if one examined the data concerning serious injuries resulting from rural road crashes, the safety situation might, in fact, be growing worse in OECD Member countries. Thus, using only fatalities as a measure for road safety tends

131 to mask the total road safety problem. Unfortunately, as this report has shown, there is scant data available to properly document all of the elements of the rural road safety problem and a full accounting of the serious injuries is therefore presently not possible.

The risk of being killed per kilometre driven on rural roads is generally higher than on urban roads and four to six times higher than on motorways. Rural road accidents are generally more severe than accidents on urban roads due to differences in operating speeds (higher on rural roads), road geometry (rural roads have evolved rather than having been designed), functionality (rural roads are multi-functional), enforcement levels (rural roads receive a lower priority) and other factors. This accounts for the increasing relative importance of rural road fatalities in relation to total road crash fatalities, which has climbed from less than 55% in 1980 to more than 60% in 1996 in the OECD Member countries. Because OECD countries have generally experienced a reduction in the total number of road crash fatalities in the same period, it is clear that motorway and urban road safety improvements have been more successful than those on rural roads.

The conclusions from these data are inescapable: the rural road safety problem is very serious and all road safety indicators (size, risk, development over time) clearly call for decision makers and the road safety community to give far more attention to rural road safety problems. From all appearances, the rural road safety problem has been neglected over the years in comparison to the high level of attention that has been given to the safety problems on motorways and urban/residential roads and streets. This is evidenced by the general lack of explicit safety policies or targets for rural roads in most OECD countries. Given this state of affairs, the rural road safety problem deserves a higher priority in future road safety policies, without neglecting the urban road safety problem.

IX.1.2 Main characteristics of the rural road safety problem

As much as 80% of all accidents on rural roads fall into three categories: single vehicle accidents -- especially running-off the road, head-on collisions and collisions at intersections. Single vehicle accidents constitute 35% or more of all fatal rural road accidents. This type of accident is the most prevalent because all three elements of the family of hazard factors -- i.e. the driver behaviour, the vehicle, and the road (infrastructure) environment -- play a role in causing these accidents and increasing their severity. Head-on collisions make up nearly 25% of all fatal accidents on rural roads. Though driver behaviour -- i.e. speed -- and the road environment -- i.e. conflict caused by the non-separation of opposing lanes -- are the principal factors in these accidents, vehicle technology has the potential to lessen the severity of the accident itself. Collisions at intersections account for about 20% of all fatal rural road accidents. Again, driver behaviour and road infrastructure are the key contributing factors to these types of accidents.

Rural road accidents are scattered over the entire rural road network, excluding some specific unsafe locations. Under these circumstances, a pressing challenge for safety professionals is to understand the causes of these rural road accidents and the contributing factors. A main conclusion from this analysis is that the rural road system itself has inherent characteristics that significantly contribute to the high number of accidents and the high risks.

Inappropriate and excessive speeds are a key factor in rural road accidents because the actual speeds on rural roads are relatively high (80 – 120 km/h) under circumstances where these high speeds cannot be driven safely all the time and everywhere. For example, rural roads generally do not have consistent design characteristics over their total length. This is especially the case for roads that are not planned but are, rather, historical roads. This requires constant speed adaptation to account for regularly

132 changing situations and circumstances that increase the opportunities for human error and lead to higher risks for accidents. Loss of control is also a major factor, accounting for 35% of the accidents on major rural roads and up to 60% of accidents on minor rural roads. These accidents are all the more serious when vehicles collide with an obstacle. Some 40% of fatalities in rural road accidents involve an obstacle. Though there are many possible causes for losing control and running-off the road, influencing inappropriate and excessive speeds together with a safe roadside design are key elements to improve rural road safety in this regard. In many countries, the consumption of alcohol, especially on weekend nights by young car drivers, is an important factor. In other countries, fatigue or medicines and drugs play an important role. Though alcohol and fatigue factors are well known and their importance thoroughly documented, there is very little information related to medicines and drugs. In spite of this, the available information does indicate that the associated road safety problems cannot be neglected. Another striking factor that has arisen is that as much as 75% of all fatal crashes in rural areas involve drivers who live in the area. This information may have important implications for future rural road safety programmes. Aside from the main areas of attention described above, certain conclusions can be made in relation to other factors relevant to the rural road safety problem. Heavy-goods vehicles and buses constitute a special problem due to the fact that these types of vehicles have a speed behaviour that is quite different from that of automobiles. This speed variation generates more instances of overtaking, which in turn can be a dangerous manoeuvre on rural roads. In addition, it is common to find slow-moving vehicles such as agricultural vehicles, mopeds and cyclists on rural roads. When traffic such as this is using the same physical space as fast-moving automobiles, high accident risks can be expected.

IX.2 Strategy to improve rural road safety

A framework for preventive measures has been developed in this report. It is based on a set of characteristics for the rural road safety problem, as follows:

• the rural road network is a very long, historically determined network, built from components -- function, road type, design and usage -- of a very different nature; • road-user behaviour can be characterised by free flow and high speeds while, at the same time, the road user is confronted with a wide diversity of circumstances as well as a lack of standardisation and road-course predictability; • the majority of accidents on rural roads fall into three categories – run-off the road; head-on and intersection -- with inappropriate speed as a dominant factor; • only limited financial means are available for maintaining and expanding the rural road network, especially in relation to the length of the network and the low traffic volumes they carry; • many different actors -- road authorities, police, etc. -- are in charge of preventive action.

Rural road safety is completely different from motorway or urban road safety and thus requires a separate management approach. Such an approach is almost non-existent in OECD countries. It is therefore recommended that every OECD Member country should develop a rural road safety improvement strategy. It is also recommended that each country should develop short-, short-/medium- and long-term programmes that are based on a sound analysis of the problems. Such plans must pay special attention to raising awareness about rural road safety both within the general public and within the organisations of all key actors -- i.e. government, peer groups and others.

133 In short-term programmes, it is advisable to develop and implement a speed management strategy in which speed-limit setting and speed enforcement (combined with publicity campaigns) are key components. Also, a trauma-management system could be installed in the short term. In the short- and medium-term programmes, traditional infrastructure measures have to be chosen that emphasize investment to improve the quality of the rural road infrastructure. It is recommended that low-cost, effective and efficient infrastructural measures are selected that preferably fit into existing road maintenance programmes. Among the infrastructure measures that are chosen, safe roadside design is a critical element. Long-term programmes should include ITS applications among other measures.

The institutional complexities in dealing with integrated road safety management require a clear identification of the role of various actors -- road administrations, police forces, the public health sector, representatives of road-user organisations, the insurance industry, the media and others -- on a regional basis. This institutional problem is fundamental due to a lack of tradition of co-operation in many regions and countries. Co-operation based on the goal of jointly conceiving and implementing an integrated road safety management programme in an atmosphere of “partnership” is the most promising approach. A leading organisation should induce and facilitate this partnership. The leading organisation should also play a central role in gathering and disseminating all relevant information, knowledge and expertise about rural road safety.

IX.2.1 Road safety measures

Throughout the chapters in this report, various safety measures are suggested that could improve rural road safety. A considerable number of these are low cost measures. Although a structural network-wide approach is required and recommended in the report, there is a clear understanding that individual low-cost measures can contribute substantially to the safety of the rural road network. Therefore, each of the following sections provides a summary of measures.

Infrastructure

Safety measures that address infrastructure offer the most plentiful opportunities for safety enhancement on rural roads and those that are low-cost and have high benefit-cost ratios have the greatest potential for widespread use. However, even though safety is understood to be an important criterion in road design, it is still too often of secondary importance. The report strongly recommends that safety should receive explicit attention at every level of the process, from the decision to build or rebuild a road to the planning and design stages, through construction and during operation and maintenance. The basis of a safe road design is a consistent, hierarchical road network, in which each road category has a particular function to fulfil. The design characteristics of a road need to be in accordance with its function and provide “positive guidance” for road users. From this standpoint, the report recommends that rural roads should be assigned a specific function rather than trying to cater to a varying mix of functions. As well, the design of the road should be consistent with the function and in accordance with the lowest functional use of the road.

It should always be remembered that the ultimate level of safety on a road depends on the consistency of the design in all of its aspects. For instance, the report notes that a series of relatively wide curves should not be followed by a very narrow one without extensive warning and/or physical speed-reducing measures. Furthermore, it must be possible to negotiate an isolated curve or the first in a series of curves at a speed which is not excessively below the speed which is maintained on the straight section preceding it. Whereas there is a general trend for accident rates to increase as a curve becomes

134 narrower, from a safety point of view the consistency between curves along the road is at least as important. On a specific note, the report suggests that using the planning process to minimise direct access to major rural roads and/or not allowing access at bends, hill crests and at or near intersections should be a minimum requirement for ensuring safe road infrastructure.

As the main rural road accident type, single vehicle run-off the road accidents occur most often on horizontal curves rather than on adjacent tangent sections. This is also the case for many head-on accidents. The report concludes that flattening horizontal curves is an effective accident reduction measure. However, reconstructing existing curves is expensive and probably only cost-effective on higher volume roads. The report thus recommends several less-expensive measures such as removal or protection of roadside hazards, flattening side slopes, improving pavement skid-resistance, increasing the superelevation, paving the shoulders and eliminating pavement edge drops. Typical low-cost measures in this regard can also include upgrading the pavement edge line and centre line in some situations, adding raised reflective pavement markers or upgrading the advance warning. Rumble devices along longitudinal sections can also be effective in reducing run-off the road accidents. The installation of roadside markings to guide drivers through a curve or a bridge are also beneficial for safety.

The report stresses the importance of forgiving roadside concepts and roadside improvements in general because they can significantly reduce the severity of accidents. There is very high potential for improving overall safety by treating or removing roadside obstacles such as trees, ditches, rocks, utility poles and steep slopes. As well, the report indicates that obstacle-free zones of between 4 and 10 meters are desirable if the road geometry and right-of-way will allow it. Finally, the report clearly identifies knowledge transfer and training in the area of roadside safety as a key action area that can contribute to better and more timely treatment of roadside hazards.

In relation to head-on collisions, the report shows that prevention can be accomplished by (physically) separating opposing traffic. A rather drastic approach that can be implemented on rural roads is narrow physical separation by means of a steel or concrete barrier. In order to reduce head-on collisions caused by overtaking manoeuvres, the provision of conflict-free overtaking opportunities -- i.e. regular overtaking lanes or climbing lane installations with good forewarning -- can have many advantages. In addition, the report shows that a combination of increasing lane width and shoulder width is the most effective approach for preventing a variety of accident types, including head-on collisions.

In considering intersection collisions, the report concludes that roundabouts have a very good safety record in comparison to three- and four-way intersections. Roundabouts should be considered for their safety record. However, because roundabouts are a relatively expensive alternative, the decision to install roundabout intersections must be based on a thorough analysis of the cost-effectiveness of this solution in comparison to others. Channelisation as a remedial measure at existing ordinary intersections can be profitable, even if the AADT is less than 7 500 vehicles. In addition, road lighting at intersections will reduce the number of night-time collisions in some conditions, though it is important that the lighting poles or masts in the roadside or median do not contribute to the number of injury accidents through poor design or location.

In addressing the issue of speed variance on rural roads, the report concludes that separating slow and fast traffic will contribute to the overall safety of rural roads. There are a number of ways to accomplish separation: i) a parallel road or secondary traffic area for all types of slow-moving vehicles; ii) a parallel, physically separated bicycle/pedestrian lane; iii) a lane at the outer side of the normal running lane for bicycle/pedestrian use only; and iv) a multi-purpose lane on the outer side of the road which in principle is assigned for bicyclists/pedestrians, but which may be used by slower-moving motor vehicles to allow faster traffic to pass.

135 As a final comment on infrastructure, the report highlights the importance of combining remedial black-spot programmes that target specific problem sites with preventive safety impact assessments and safety audits, as appropriate, when planning, designing, (re)building or maintaining roads. The aim is to prevent accidents rather than respond to those that have already happened. The report therefore recommends widespread use of these practices at the local, regional and state level.

Police enforcement

The report found that police enforcement is an effective symbol to show that road safety is as important as other types of crime and misdemeanour. This is especially important given the contribution of inappropriate speed and excessive speed in rural road crashes. Effective enforcement can serve as a general deterrent factor that can bring about long-term behaviour changes in drivers if it is coupled with other firm actions including appropriate penalties and sufficient driver training. However, due to the great length of the network, enforcement by conventional means is very limited and one cannot rely only on strategies based on “improving behaviour on the spot” by spending police manpower alongside the road. The report concludes that publicity campaigns associated with targeted enforcement can increase the enforcement effects and contribute to a change in driving norms. In a similar vein, the report concludes that repeated enforcement creates longer halo effects, in terms of both time and distance, in contrast to “blitz” campaigns. By introducing a random enforcement element, enforcement effectiveness can also be increased and longer halo effects will be produced. The report recommends that automated enforcement technologies that target the causes of the principal rural road accidents should be considered. Finally, the report strongly recommends that funds generated by traffic enforcement be earmarked for rural road safety to ensure that these important safety problems are addressed to the fullest extent possible.

Intelligent Transport Systems (ITS)

The report concludes that the full potential of ITS solutions for rural road safety can be realised only if research is undertaken to obtain a better understanding of the costs of these systems, the specific technical issues, the human-machine interface, and the institutional and political constraints. As well, the extensive nature of the rural road network demands low-cost solutions. In spite of this cautionary view, the report does identify a host of low-cost ITS measures that will be ready for deployment within the next three years and that could contribute to reducing the principal accident types on rural roads. Paramount among these, given the major role of speed in rural road accidents, are speed-control technologies such as speed advisory systems and adaptive cruise control. Other near-term, low-cost measures include systems for driver monitoring, intersection approach warning and guide-lights. In the next three to seven years, other low-cost measures such as smart seat-belts and air bags or vehicle data recorders will be broadly available and can lessen the rural road safety problem. Finally, the report identifies ITS measures that are high cost and/or will not be available for some time. Decisions to apply these measures in rural road situations must be made on a case-by-case basis.

Trauma management

Identifying an accident location was pointed out in the report as one of the key problems in responding to rural road crashes. The report cites several options that can improve the situation, including: improving road and kilometre/mile identification schemes; expanding the use of GPS; and exploring possibilities for automated accident detection. Several communications technologies should also contribute to improving rural road safety. Among available technologies, cellular telephones are

136 viewed as an extremely positive advance as they can shorten arrival time and improve the overall information available about an accident situation. The report suggests a role for publicity campaigns in conjunction with more widespread first-aid training to improve trauma treatment at the scene of a rural road accident. The report also recommends and describes common guidelines and standard procedures that local hospitals could adopt to improve trauma treatment.

IX.3 Research needs

The following sections describe rural road safety research needs that were identified as a result of the work of the Expert Group. It should be noted that each chapter in the report includes very specific research recommendations whereas this section provides an overview of the primary research needs that emerged.

IX.3.1 Data collection and analyses

One of the significant conclusions reached by the Group is that there is currently insufficient information available on rural road safety problems to adequately support appropriate policy and investment decisions. This is important because improving rural road safety will require standardised methods for collecting and reporting accident data, identifying exposure measures, monitoring and evaluating countermeasures and estimating the cost effectiveness and benefit-cost ratios of these countermeasures. With these standardised methods in place it would be possible to build a sound basis for rational rural road safety policies. Therefore, more systematic evaluation of the effectiveness of countermeasures is necessary based on valid and reliable data. In this regard, benchmarking of rural road strategies may help to improve effectiveness. It is recommended that more international exchange of research results be encouraged -- perhaps utilising the International Road Research Documentation (IRRD) database -- and that an international platform to deal with programming and executing meta-analyses on important topics of rural road safety be created. This platform should have the means to make rural road safety knowledge easily accessible world wide using modern dissemination methods and techniques.

IX.3.2 Rural road safety strategies

The strategic framework developed in this study needs more research in order to establish the benefits associated with this approach. The key elements to be researched are: i) mono-functionality of roads in the rural road network and the consequences for road design; ii) methods to explicitly deal with road safety when improving or maintaining the rural road network; and iii) developing an integral approach to improve rural road safety. It is recommended that possible pilot projects be identified that address these areas in conjunction with an examination of the institutional arrangements between stakeholders.

IX.3.3 Rural road safety measures

It is quite evident that the current knowledge and expertise about how to improve rural road safety is not sufficient. This lack of knowledge is found in various examples throughout the report. For instance, there is insufficient understanding about why road users make errors that sometimes lead to accidents or why, on a massive scale, they do not obey speed limits. Knowledge is also rather limited

137 regarding how to influence human behaviour effectively and efficiently. The development of comprehensive research programmes to fill these knowledge gaps is therefore recommended. An important element in the development of these research programmes is the use of certain road safety theories or models for guidance. In addition, knowledge transfer can be maximised by giving explicit attention to safety as a basic design element in university-level road engineering courses. Moreover, a vision or strategy for improving road safety -- e.g. the sustainable safety vision in the Netherlands and the zero-vision in Sweden -- could be regarded as extremely helpful in this regard.

IX.4 Next steps -- dissemination of the results

A broad dissemination of the results of this study is justified and advisable given their economic and policy relevance. This recommendation is strengthened by the results of the analysis of the traffic risks (or dangers) on rural roads that leads indisputably to the conclusion that OECD Member countries must pay more attention to the safety problem on this part of their road networks. The Group also recommends that the safety strategy concept developed in this work be broadly implemented in the Member countries.

In order to advance these recommendations, it is necessary, as a first step, to bring them to the attention of key stakeholders in OECD Member countries, particularly the national organisations responsible for road safety policy. Secondly, they should be promoted with the supranational/international organisations -- e.g. World Bank, Regional Development Banks, PIARC, IRF, APEC, REAAA and others -- that are most likely to be interested in taking further steps in policy and research based on the conclusions and recommendations. Thirdly, they should be marketed outside the OECD Member countries where there could be a high level of interest in the results. Though all of the contributing countries share responsibility in these three levels of dissemination, initiative on the part of the Road Transport Research Programme of the OECD is also required to assure the fullest dissemination of the results.

It is quite clear that this report contains a profusion of information based on a wide variety of international experience on ways to advance rural road safety by improving the planning, design, maintenance and operation of road infrastructure, by modifying driver behaviour through better police enforcement and information campaigns, through the targeted application of ITS and by improving trauma-management systems on rural roads. It is therefore recommended that an international symposium be organised where the results of this study can be presented. It is also recommended that the countries participating in the RTR promote the dissemination of these results in professional journals and through presentations at congresses and seminars. Furthermore, it is advised that this report be brought to the attention of international organisations with a view to giving them the opportunity to react on the results of this study with their own activities.

138 Annex A

LIST OF MEMBERS

Chairman: Mr. Wegman (Netherlands)

Australia Mr. Denny O’LEARY

Belgium Ms. Jan PELCKMANS

Canada Mr. Randolph W. SANDERSON

Czech Republic Mr. Jan SPOUSTA Mr. Josef MIKULIK

Denmark Mr. Per MATHIASEN

Finland Ms. Saara TOIVONEN

France Mr. Christian MACHU Mr. Thierry BRENAC Mr. Dominique FLEURY

Japan Mr. Tamotu KOBAYASI Mr. Masakazu NAKAGAWA

Netherlands Mr. Fred WEGMAN Ms. Ingrid van SCHAGEN Mr. Joop H. KRAAY

Switzerland Mr. Uwe EWERT

United Kingdom Mr. Adrian WADDAMS

United States Mr. A. George OSTENSEN

Corresponding Mr. Shalom HAKKERT, Israel members Mr. Peter HOLLÓ, Hungary

OECD/RTR Mr. Wolfgang HÜBNER Mr. Anthony OCKWELL Mr. Patrick HASSON Ms. Véronique FEYPELL- de LA BEAUMELLE

139 Contributing authors

Mr. O’Leary, Mr. Wegman, Ms. Van Schagen, Mr. Ewert, Mr. Ostensen, Mr. Hakkert, Mr. Hasson.

Editorial committee

Mr. Spousta, Mr. Wegman, Mr. Waddams, Mr. Hakkert, Mr. Hasson, Ms. Feypell-de La Beaumelle.

140 OECD PUBLICATIONS, 2, rue AndrÂe-Pascal, 75775 PARIS CEDEX 16 PRINTED IN FRANCE (77 1999 01 1 P) ISBN 9264-17054-5 ± No. 50643 1999