Incident Management Study, Issued July 2002

Acknowledgements

Executive Summary

  1. Introduction
  2. Methodology
  3. A Better Understanding of Incident Management
  4. Review of Incident Management in the US
  5. Main Analysis
  6. Conclusions
  7. Recommendations

List of Appendices

  1. The Four Incidents
  2. Part 1: Typically What Happens & Part 2: Incident Response Targets and Frequencies (anecdotal)
  3. USA Experience - Costs and Benefits

List of Figures

  1. Flow Chart of Activities
  2. Estimate of 'Lost' Vehicle Hours using Default Values in INCA
  3. Timeline of Incident Management Activity
  4. A Framework for Organising and Sustaining Incident Management Programmes
  5. Congestion Benefits / Costs
  6. Model of Incident Management Tasks
  7. Timeline of Incident Management Tasks

List of Tables

  1. Analysis of Errors
  2. Results of Task Analysis
Incident Management Study, Issued July 2002

Acknowledgements

The authors would like to acknowledge and thank the following organisations that greatly assisted the study team during the course of this work:

Participants in Workshops

West Midlands (CMPG)
North Yorkshire
Cheshire
Site visits and in-vehicle observations
Acknowledgements

Executive Summary

Purpose of Study

The purpose of the study was to gain a better understanding of what happens during major incident clearance; the roles and responsibilities of those organisations taking part; and to provide recommendations for improvement.

Context of Study

Good incident management is a very important goal for all concerned with the management and safe operation of the road network. Incidents account for about 25% of congestion on the trunk road network. Incident-related traffic congestion (including secondary impacts) detrimentally affects public safety, the local economy and the environment.

The management of the trunk road network has changed significantly over the last 5 years or so with the introduction of new arrangements for managing this. New initiatives are being pursued but often on an area by area basis. New technology also brings opportunities to improve the efficiency of incident management but which will necessitate closer liaison between the police and highway authorities. Against this background, the Motorists' Forum has questioned whether it was possible to reduce the time taken to clear incidents with the suggestion that a performance indicator be agreed to halve by 2004 the average time taken to clear the effects of serious incidents.

Study Method

The study method was essentially qualitative in nature and covered three main activities:

  1. Gathering of data and information to provide a greater depth of understanding of current practices in the UK and experiences elsewhere (literature review, meetings and site visits).
  2. Workshops in three locations (West Midlands, Cheshire and North Yorkshire) to identify current practices, the sequencing of actions and interactions between the organisations involved and key issues and areas for improvement.
  3. A formal qualitative task analysis of data gathered.

Main Conclusions

  1. There are a number of relatively short-term but potentially high profile initiatives that can be taken to aid incident management such as enhanced location signing on featureless highways and permanent symbol signing of diversion routes. These will clearly have a resource implication. Initiatives that emerge as most needed, in relation to the issues identified in the workshops, are those that concentrate on foreshortening the 'platinum' period i.e. the first 20 to 30 minutes after the incident is reported.
  2. The setting up of targets for major incidents is hampered by the complexity of such incidents and the fact that no two are the same. Mechanisms for evaluation can and should be developed on the basis of local agreement between all the key organisations. Participants at the workshops were keen to improve their own organisations' performances through greater training, management and co-ordination with others. However, the imposition of targets from above would not at this time achieve wider ownership of integrated incident management processes. Incident management should involve all the key stakeholders to maximise the potential of new technology.
  3. Very different highway operating environments apply which work against the setting up of an overall national target e.g. to cut clearance time by half. Technological innovation in the form of new management systems, such the traffic control centre, is expected to help reduce overall incident clearance times. This could come about through the use of new technology, particularly with regard to incident detection and response, and greater integration between the police and highway operators. There is currently a dearth of data relating to incident clearance which needs to be addressed.
  4. There is a distinction between major incidents, where there is an interaction between a number of organisations, and less significant ones which may disrupt the flow of traffic but not involve personal injury. The Highways Agency (HA) is pursuing initiatives in these areas with the advent of Rapid Reaction teams in Area 8, 10 and 16 (vans and motorcyclists) and Minuteman (to clear light vehicles) on the A1 and A63. Both have achieved demonstrable benefits in economic terms. However they need to be seen as part of a wider area strategy, all types of incident, which is integrated with those organisations responsible for incident clearance.
  5. No one organisation has overall responsibility for 'Incident Management' - each plays its own part and may pursue a wide variety of initiatives aimed at improving this. Informal agreements and liaison do take place between the various organisations - often at middle management level and as a result of major incidents. However the development of sustainable cross-organisational agreements and plans (including the setting of targets and monitoring) would be severely hampered by the lack of more formalised incident management processes and organisational clarity.
  6. The HA is the only organisation involved with incident clearance that has congestion as a key performance indicator. The benefits of incident management leading to reduced clearance times are, in large part, measured in terms of improved traffic flow and reduced congestion.
  7. The trunk road network should be viewed as a 'safety critical system' in the same way that the railway network is one. This is an important metaphor to take on board in planning and development of the network. The reason for this is that it changes the way the roles of all key stakeholders might be viewed. There is a range of established methods for training people and developing organisations which emerge from best practice in the management of safety critical systems. These provide a number of ways forward for the management of the trunk road network including a recognition that the traffic police are, presently, the operators of the system in a very real manner. As such they undertake an essential role in underpinning incident management which cannot be replaced by less trained personnel. The Home Office's 'Police Reform' programme may however present opportunities for a limited transfer to accredited organisations/wardens engaged by the HA to assist with incident management.

Recommendations - Short Version

Recommendation 1 - Operational
Short term but potentially high profile initiatives should be reviewed with a view to implementation. These include enhanced location signing on featureless highways and permanent symbol signed diversion routes.

Recommendation 2 - Institutional
The Highways Agency should take on the key role of developing, negotiating, implementing and monitoring better incident management procedures.

Recommendation 3 - Monitoring
An incident database should be established, initially in trial areas. The concept of incident recording should then be rolled out to other areas or regions. Regional agreements on evaluation measures can then be developed from local incident data but co-ordinated at a national level.

Recommendation 4 - Organisational
A top-down review should be undertaken of such organisational factors as contractual arrangements, cross-organisational working practices, management practices and allocation of responsibilities at all levels. This review to include a proper assessment of the possibilities of using non-police management of some incidents or elements of incidents including accident investigation.

Recommendation 5 - Guidelines and Planning Support
The lead body, jointly with ACPO, should develop the output from this study (with other stakeholders where appropriate) to produce an agreed 'National Guidance Framework' (NGF) for the top level 'good practice' incident management culture, procedures and processes.

Recommendation 6 - Guidelines and Planning Support
Thereafter the NGF to inform all service providers within a region which, whilst reflecting the tenets of the NGF, enables unique Detailed Local Operating Agreements (DLOAs) to meet the topographical, institutional and infrastructural needs of the area.

Recommendation 7 - Training
Linked with Recommendations 5 and 6, a pool of training and development materials should be prepared, cross-organisational training values and priorities developed and other training models related to management of safety critical systems reviewed.

Recommendation 8 - Technological
Longer term technological initiatives which can assist incident management should be critically reviewed. These include improvements in communication systems, camera equipment, GIS/GPS and transfer of CCTV images to other emergency services and clear-up contractors.

Executive Summary

1. Introduction

1.1 Scope of Study

URS Corporation Ltd, in association with University of Hertfordshire and ITS, Leeds University, were commissioned by the Highways Agency, on behalf of the Incident Management Group of the Motorists' Forum, to undertake a study into the management of incidents. The aim of the study was to gain a greater understanding of the way in which unplanned incidents are dealt with on dual carriageway roads; the roles and responsibilities of those involved; and the key factors involved. Through this understanding, it should be possible to identify improvements that can make incident clearance procedures more effective in today's operational environment.

This report describes the background to the study; the approach taken to explore the above issues; and the main findings and recommendations that have emerged from the work. Further detailed reporting on the analysis is contained in a separate volume of appendices which also contains detailed report summaries from the workshops and other interviews. Copies of these can be obtained from the Highways Agency contact address set out on the back cover.

1.2 Background

Good incident management is a very important goal for all concerned with the management and safe operation of the road network. Incidents account for about 25% of congestion the trunk road network. Incident-related traffic congestion (including secondary impacts) detrimentally affects public safety, the local economy and the environment.

Unplanned incidents on the trunk road network can range from minor vehicle spillages to multiple injury accidents. The reaction and follow-up action required for these will clearly vary according to the nature and extent of the incident as will the roles of the various bodies responsible for dealing with it.

The management of the trunk road network has changed significantly over the last five years with the creation of area agents responsible for maintenance over specific geographic areas. The contractual arrangements within each of these can vary depending on each individual contracts. New management initiatives are being pursued but often on an area by area basis. These bring into sharper focus the various roles and responsibilities of those responsible for managing the road space before, during and after the incident and attendant emergency and other services.

Against this background, the Motorists' Forum has questioned whether it was possible to reduce the time taken to clear incidents on the primary route network with the suggestion that a performance indicator should be agreed to halve by 2004 the average time taken to clear the effect of serious incidents. Currently, a fatal incident or a serious injury that might become fatal is expected to involve total or partial road closure for a significant time to allow police sufficient time to investigate. This can, on certain parts of the network, represent a substantial cost in terms of delay.

There are clearly gains to be made in developing an effective incident management strategy involving multi-agency responses to traffic disruption that result in congestion. However care needs to be taken in involving all parties so that wider ownership of proposals is achieved. A solid basis of understanding together with factual data and information on incident clearance needs to be established before monitoring regimes can be realistically considered. In this way, new processes which are designed to improve incident management become more accepted and can be used to as an objective tool of assessment.

1.3 Objectives

The specific objectives of the study were to:

  1. Identify all the organisations involved in handling incidents; establish their precise contribution to the exercise; and show how the performance of those roles affect the overall clearance of incidents.
  2. Ascertain what information is available (statistical and anecdotal) on the management and speed of clearing incidents.
  3. Identify examples of 'good practice'.

1.4 Methodology

The study methodology was essentially qualitative in nature and covered three main stages:

  1. Gathering data and information that provided a greater depth of understanding of current practices in the UK and experiences elsewhere (literature review and meetings).
  2. Workshops to identify current practices, the sequencing of actions and interactions with other organisations.
  3. Formal qualitative analysis of information gathered.

The flow chart in Figure 1 at the end of Section 2 depicts the main study stages. The emerging findings report was produced in February 2002.

1.5 Report Structure

Section 2 of the report describes the methodology adopted whilst section 3 seeks to provide a better understanding of incident management as defined by the workshops and earlier meetings and site visits. Section 4 provides a review of the US experience in this field whilst the main task analysis is described in Section 5. Conclusions and recommendations are provided in Sections 5 and 6.

1. Introduction

2. Methodology

2.1 Initial Review

The initial tasks identified in Figure 1 involved the gathering of data and information in order to provide the study team with a greater understanding of current practices in the UK and elsewhere.

In fact very little literature was available to describe UK experiences other than descriptions of the HA and maintenance agency initiatives including 'Minuteman' and 'Rapid Reaction' teams in specific areas. A lot of literature was available from the US where incident management is far more formalised in terms of roles, responsibilities and management. Reference will be made later to this.

In addition to the literature reviews, meetings were held with a number of representatives from the key organisations involved in incident management. These included:

Group interview sessions were also held with members of Hertfordshire Constabulary on 16 November 2001, with follow-up sessions going out on patrol with the police to incidents on 30 November 2001 and 18 December 2001.

2.2 Workshops

Information gathered at the earlier stages helped to provide a firm basis upon which to plan the workshop sessions. These were the primary source of data collection and were designed to:

It was originally envisaged that workshops would be carried out at the following locations:

However, in parallel to the work being undertaken for this study, a similar workshop was being organised by the HA encompassing the M25 Sphere in December 2001. The need for this was triggered by a significant chemical spillage incident on the M25 resulting in substantial delays and so the workshop was focused upon improving incident communication procedures specifically within the M25 region.

To avoid duplication, workshops for this study were held in the other three locations. It also became apparent from these that sufficient information had been obtained to achieve the study objectives. The M25 Sphere workshop also focused more on the maintenance agency aspects of incident clearance, which provided a good balance with the three other workshops, which were strongly represented by members of the emergency services. The results from the M25 Sphere workshop were provided by the organisers and are included in this report where relevant. The remaining three workshops were held on:

Attendees at the three workshops were drawn from the following organisations:

2.3 Design of Workshops

The workshop delegates were organised into a number of groups consisting of a range of different organisation representatives. For the final workshop in Cheshire, for the afternoon session the groups were selected to be homogeneous in terms of organisation representatives. Central to the workshops were discussions around example major incidents or vignettes. These were as follows:

The incidents were shown on AO sized computer generated graphic displays along with associated assumptions regarding, for example, the number and types of casualties, road type environment, the nearest junction for access, proximity of houses, wind direction, other distractions e.g. animals. The incident descriptions are set out in Appendix A.

The vignettes provided a central theme to the workshops around which roles and responsibilities could be discussed, key tasks identified and problems and solutions examined. A simplified schedule of activities for each workshop is provided below. Further detail can be found in the workshop reports that provide timelines for each group and other information on key tasks, problems and solutions.

2.4 Analysis

The analysis was undertaken in two main parts. Firstly, the basic information was summarised, under the following headings:

Secondly, the data gathered during the workshops, interviews and site visits were used to break down and analyse the tasks performed by each organisation in detail.

Section 3 describes the basic information obtained from the workshops whilst Section 5 describes the key findings from the task analysis.

Figure 1

Figure 1 (20KB PDF)
2. Methodology

3. A Better Understanding of Incident Management

3.1 Incident Types and Performance Data

The starting point for any interpretation of what actually happens during incident management is an understanding of the general time line of activities and the roles and responsibilities that those involved perform. A clear message from the workshops however was that no two major incidents are ever, or very rarely, the same and that an endless number of consequences may take place depending on local circumstances. Many factors affect the characteristics of an incident including:

It is also worth noting that most incidents are minor and require police attendance only. It is reported that fire and rescue services (FRS) attend 3% of all incidents and the ambulance service somewhat less. In one county, this may equate to less than 10 RTI's (road traffic incidents) per month for FRS. Much analysis concentrates on the 3% of major incidents.

Research work for the former Department of Transport, Local Government and the Regions identified generic incident frequencies and durations by incident type in order to identify economic benefits resulting from road improvements using software called INCA (INcident Cost benefit Assessment). The amount of time lost per year at different flows measured in AADT (Annual Average Daily Traffic) can be calculated using INCA depending on the flow in each flow group, percentage HGV and lane capacities. Using default values in INCA, Figure 2 provides an indication of 'lost' time (vehicle hours) per year on a 1km stretch of motorway for different types of incident and for different flow ranges.

Subsequent work undertaken on behalf of HA on specific routes indicates that the assumptions in INCA concerning incident frequency and durations for a road with a hardshoulder (motorway) are of the right order of magnitude accepting the problems alluded to above concerning incident definition.

View larger image

Figure 2: Estimate of lost vehicle hours using default values in INCA

It is worth noting here that no systemised or standardised method of incident reporting is available across the UK and that, thus far, reporting of incident frequencies and durations has been undertaken on an ad-hoc basis to address specific issues. This involves an assessment of police (detailed) or vehicle recovery organisations (far less detailed) records to evaluate the performance of e.g. new signing systems, assess new methods of dealing with minor breakdown clearance (HA's Minuteman) or providing a basis for calculating the algorithms in INCA. In some cases, incident reports are only maintained for a year and then discarded by the holding authority. The motoring organisations also maintain incident data which can be used by members to help plan journeys e.g. via the internet and which can also be used to provide regional and global summaries of incident numbers (as reported) and to identify listings of 'hot-spot' locations.

3.2 Roles and Responsibilities

While there are common elements in many 'clearance' tasks, no two such tasks are the same. Accepting this however a generic timeline is presented as Figure 3. This attempts to combine and simplify the various timeline of activities identified by the participants at the workshops for a range of sample incidents. These involve major incidents of the type presented to workshop participants where attendant emergency services would be involved.

Typically, for the larger scale incidents discussed at the workshops, the following are key participants:

Incident Clearance Accident investigations following fatal incidents can take some time to perform. Specialist equipment is available on the market and it is claimed that this is capable of significantly reducing this time. Specialist heavy lifting equipment is available on the market capable of significantly reducing the time taken to right overturned HGVs. Motorists' information It is vital to get accurate and timely information out to the motoring public in areas a long distance upstream of an incident. There are many media outlets of information to the public, with a variety of methods for obtaining the information to output and the quality of that information.

The Police:

The police are often the first to receive notice of an incident (as receiver of '999' calls) or the first to detect incidents because of their role in traffic patrol and traffic law enforcement. They are typically in command at the scene, request additional services, and lead crash investigations when the incident results in personal injuries, fatalities, or significant property damage. The police have the key role in incident management and associated tasks which are described in more detail in the next section.

Fire and Rescue Services:

These include county or city fire and rescue organisations. They respond to incidents involving fire, hazardous materials, medical emergencies, life support or rescue, and thus play an important role in incident management. Fire and rescue services do not have a statutory duty to attend motor vehicle accidents but do enjoy statutory 'control' powers at a 'fire ground'.

Ambulance and Paramedic Services:

These provide an essential service to those injured in road traffic accidents. Typically their involvement is at the early stages of a major incident and they have left the scene well before the road is reopened to traffic.

Transportation Agencies:

These include departments of transportation and other agencies that operate and maintain the road network. The Highways Agency is responsible for the trunk road network whilst responsibility is given to local highway authorities on non-trunk roads. Clearly incidents occurring on trunk roads have wider impacts than on the trunk road link itself and representatives from both attended the workshops. HA's trunk road network is divided up into a series of Areas which are managed on a contractual basis by maintenance and contractor agents. These agencies generally provide traffic management support, incident information dissemination to other impacted organisations, equipment and personnel for incident clearance, special signing, maintenance of sign 'dumps', activation of detours, containment of minor hazardous material spills, debris removal and related activities.

Hazardous Materials Cleanup Services:

These services are generally provided by specialist companies such as BIFFA. Significant spills may involve not only the HAZMAT service but also related environmental protection authorities from the Environment Agency.

Towing and Recovery Companies:

These are private companies that provide towing and recovery services for highway incidents. They are often under contract to one of the agencies involved in incident management, may independently patrol the highway, or be contracted by the motorist. Specialist recovery resources, such as heavy lift or rotator vehicles, are usually privately contracted by an involved agency.

Information and Private Traveller Information Providers:

These include public agencies and private companies such as information service providers that collect, process and disseminate traffic and transport related information to benefit travellers. Common methods to disseminate information are television, radio, the Internet, highway advisory radio, and variable message signing.

The magnitude or nature of an incident may require other organisations to respond or participate in incident management on an as-needed basis.

The workshops brought into sharp focus the differences in work culture between the different organisations. In particular:

No single organisation is statutorily responsible for incident management in a holistic sense although the police do enjoy primacy at the scene. Each plays its own part. Initiatives pursued by the HA such as new Rapid Response Vehicles (e.g. Area 10, Area 8) and Minuteman (A1, A63) also bring into sharper focus the roles and responsibilities played by each, as does the advent of new road management systems such as MDIS (Midlands Driver Information System) and the TCC (Traffic Control Centre).

The following identify the phasing of a large incident though again, the precise details of each will vary greatly depending on local circumstances:

3.3 Key Issues

Introduction

A number of key issues arose out of the workshops and these are described below. As well as the key issues, a lot of factual information was provided at the workshops which helped to provide a much better understanding of the complexity of incidents and what actually goes on in terms of the lines of communication and actions taken by key participants. This information is provided in Appendix B - Part 1 along with a list of target response rates and incident frequencies (anecdotal) in Part 2.

Roles and responsibilities

There is a lack of definition about the roles and responsibilities of each of the participants involved in incident management and a lack of national guidelines and procedures for incident management. It was stressed at all of the workshops that cross-organisational training and debrief sessions would help to clarify roles and responsibilities and highlight areas for improvement. At the moment, many organisations do not have practice exercises and very few organisations hold even in-house debrief sessions. Contingency planning was an issue raised at two of the workshops, with the suggestion that set procedures for dealing with different incidents in different areas could be developed and combined with tabletop exercises.

Non-hazardous spillages may be the responsibility of the term maintenance contractor or the vehicle recovery agent, but it is not clear which organisation has responsibility for clearing up medical waste. Similarly, it is not clear who has responsibility for holding and maintaining drainage plans that would show the location of balancing ponds and other pollution control measures. At the workshops it was stated that on occasions the Environment Agency has the plans but the Highways Agency does not. At other times, the Environment Agency may have to buy them from third party consultants.

Training

Some of the representatives at the workshops, and those interviewed individually, felt that training to undertake site management tasks was inadequate. This was substantiated by earlier interviews. The incident scene can be a complex and dangerous environment. It exposes both victims and those responding to the incident to any combination of fast moving traffic, hazardous materials, fire and electrical hazards, damaged vehicles and stressful weather conditions. With the advent of the Emergency Response Units or Teams, there is the possibility that these services may be the first on the scene. There was concern that the personnel manning these vehicles might not be trained to make the initial assessments, or to cope with severely injured casualties. The priority for the emergency services is to preserve life. The police do not advise any of the emergency services to attend to casualties; vehicle fires or spillages until temporary emergency road or lane closures and signing are in place because of the risk to personnel from passing traffic. In North Yorkshire, where the police resources were considered to be severely constrained, it can often be the case that road closures and signing may not be dealt with promptly and efficiently and only consist of a police vehicle placed in the fend-off position. If there is only one officer on the scene he cannot perform the initial assessment and set up cones and signs simultaneously on his own.

Pinpointing the location

The majority of incidents are brought to the attention of the emergency services by means of telephone calls made by the travelling public, either via the emergency roadside telephones (ERT's) or by personal mobile cell-phones. Calls received via the ERTs are linked to the referencing system by which the location of the ERT is determined. The location of the ERT will then provide the emergency services with information about the road and the carriageway affected by the incident. Calls received via personal mobile phones, however, are not linked to any location referencing system that the emergency services are able to access. Therefore detailed information about the road and carriageway affected by the incident will have to be extracted from the caller.

People often do not keep track of their location to the level of detail the emergency services require. The problem is worse on motorways than all-purpose dual carriageways because link lengths are generally longer, often featureless and junctions further apart. Delegates at one workshop estimated that up to 10% of calls received by British Telecom on the '999' number from a mobile phone were directed to the wrong area's Control Room. The public is generally unaware of the road marker system - the Control Room often has to explain what the marker posts are to the caller so that the reference post can then be identified by the caller. The problem can be eased or compounded by the proliferation of calls from mobile phones which can either verify or conflict with the location given by the first caller.

Boundary problems

The contractual arrangements for the managing agents and term maintenance contractors from one Highways Agency Area may preclude their involvement in another adjacent area, even if their resources are the closest to the scene. This can cause considerable delays to incident clearance particularly if the managing agent / term maintenance contractor needs confirmation from the Highways Agency that the loan of equipment/resources will not lead to failure in meeting their own performance targets.

As well as ensuring the appropriate response time, organisations need to ensure that they provide the appropriate response in terms of personnel and equipment. Most of the organisations involved in incident clearance do not follow the same boundaries as the emergency services which themselves may follow slightly different boundaries or cover more than one county. Resources may often be needed in other areas than the police boundary in which an incident has occurred and allowance needs to be made for this fact to provide the non-emergency services additional time to gather the appropriate personnel and equipment. Hence, the earlier these organisations are informed of the incident, the more likely it is that they will be able to supply the required resources efficiently when needed on site.

Too many cooks!

Conversely, there is no need for these services to actually be on site during the emergency phase of an incident - this could result in scene congestion, and in any case the requirements from that service at the scene could change during the course of the incident and different or additional personnel and equipment may be needed. This is a conflict that all of the workshops highlighted as a major problem and one which needs resolution if incident clearance is to improve. There is also the risk of the dilution of the quality of information if there is a long communication chain to the people who need that information.

Accessing the site

For the emergency services, all will use the hardshoulder, where provided, to reach the scene of an incident. Where there is no hardshoulder, most will try to part the traffic in the middle of the running lanes. Non-emergency services do not carry blue lights on their vehicles. All of the workshops have identified this fact as a problem when trying to reach the scene where no hardshoulder is present. In cases where there is narrow or no hardshoulder, the non-blue light vehicles will have to try to 'battle' through the tailback to get to the incident, especially if a police escort is not available - the orange / yellow lights do not infer an emergency situation and the traffic will often not make way for these vehicles. Non-emergency service drivers do not generally undergo emergency driving training, which could make driving through congested traffic more dangerous.

Ownership

Ownership of the scene was an issue raised - the senior police officer present is in charge of the incident, but it is not always clear to the non-emergency services exactly who that person is. The enhancement of the role of Incident Commander, who remains on the scene throughout, was suggested though others questioned the practicality of such a role. This would be somebody who has the responsibilities of organising police escorts for non-emergency services, liaising with all the organisations regarding the scene requirements and acting as one point of contact, at the scene, that people can use to keep up to date with the incident progress. Sometimes communicating between the emergency services at the scene can be difficult as they all operate different communication systems, and so even when two representatives may be 500 metres apart, communications between them may have to go through the control rooms.

Finding out about the ownership of the road was also raised as an issue at the workshops. This can take a long time, especially where trunk roads are being de-trunked, or where there is boundary confusion. The police control room needs to identify the correct organisation in order that road closures/diversions etc. can be implemented. There can be boundary confusion at the borders between the roads that form part of the Highways Agency responsibility and roads that form part of the local authority responsibility. Motorways and slip roads are the responsibility of the Highways Agency whilst roads passing under or over the motorways may be the responsibility of the Local Authority. There are certain junctions on the motorway network, which form the boundary of many Highways Agency areas. For example, Junction 24 of the M1 links the M1 with the A50, the A453 and the A6. The M1 north of Junction 24 falls into Area 14, the M1 south of Junction 24, and the A6, fall into Area 11, the A453 is in Area 7 and the M50 is a DBFO road covering Area 28. Also, contractual arrangements for trunk road maintenance change every 5 years or so which can in turn lead to further confusion and 'settling-in' periods when new management systems are introduced.

Diversion Routes

Diversion routes can take a considerable amount of time to set up by hand, and the initial planning and consultative phase between the police, the managing agent and the local authority as to the route for the diversion can also be quite time consuming. In some parts of the country, strategic diversion routes are planned or are already in operation, whereby permanent diversion symbol signing is set up and all that is required for the diversion is to place signing at the beginning of the diversion. Planning diversions off motorways where there is an adjacent A-road that was the main route before the motorway was built is generally quite straightforward. Diversions off dual carriageways and other motorways can be more difficult. For example, the infrastructure in North Yorkshire was considered to be relatively sparse and if incidents affect both the A1 and the A19, planned or unplanned, there are no other roads in the county capable of taking the traffic and extreme congestion results.

Vehicle Recovery
Lorries:

The vehicle recovery agents present at the workshops have stressed the importance of their early presence at the scene to professionally assess requirements. For example the specialist vehicle recovery agent at the scenes may decide that it is necessary to close the opposite carriageway to upright a HGV, as they would need to put vehicles on both sides of the HGV to stop it from rolling right over again. If the decision to close the opposite carriageway is left to the latter part of the incident, the resources required to undertake this task may have left the scene and there would be additional delay to clearance. There is new specialist heavy lifting equipment available that enables very rapid vehicle recovery called rotator systems, but these are very expensive (about £200,000) and consequently there are only 4 to 5 of these in use throughout the country.

Contractual agreements can impact on operational decision making e.g. specialist equipment must be paid for whether or not it is actually used at the site. The vehicle recovery company can not however recover the operating cost of the vehicle if it is not used on the scene.

Light vehicles:

A proportion of the recovery work undertaken by an operator may be unpaid. When the vehicle recovery agents remove a vehicle from the scene, as they are required to do in order to remain on the list of approved operators; they have to try to recover the costs from the owner of the vehicle. As the vehicle recovery agents deal with the car owners who are not members of the AA, RAC or other such organisations, they are more likely to not be able to afford to pay for the recovery of their vehicle. In addition to the recovery of vehicles involved in incidents, the recovery agents are also required to remove abandoned vehicles, for which the owners can often not be traced, and are therefore also non-paying jobs. The recovery agents often have to pay an administration fee of between £12 to £15 per call-out in order to remain on the list of approved operators, for both police contracts and AA contracts for vehicle recovery. This fee also has to be recovered from the owner of the vehicle, or borne by the recovery agent. It has been expressed that these mounting costs will lead to the reduction in the number of vehicle recovery agents, and in the quality of service that can be afforded. It will also be likely that very few recovery agents will be able to afford the new heavy lifting equipment that could help to speed up incident clearance significantly.

Motorists' Information

The workshops have indicated that it is vital to get accurate information output by the media in areas a long distance upstream of the incident. This would benefit both the motorists and the organisations involved in incident clearance, because the motorists would not get caught up in the traffic queue, making it easier for the incident clearance teams to reach the incident. It is apparent that a lot of harm can be done when inaccurate information discredits the system - if a matrix sign displays a 40mph speed limit when an incident has been long cleared, motorists are more likely to ignore speed limit signs in the future.

Very few of the outlets for information cover the period in which the incident has been cleared though, and so motorists may believe that there are still delays and divert down less suitable roads, when in fact the motorway may be clear.

It was suggested that if motorists see a variable message sign informing them of major delays up ahead, 30% of them will divert, and if they hear an announcement on the radio 30% of them will divert, but if they see a variable message sign and hear an announcement, 70% of them will divert. The importance of getting accurate and timely information out to the public was stressed at all of the workshops, yet the task of informing the media comes quite far down on the task list.

Summary of Key Issues / Problems
Response
Site Management
Traffic Management
Incident Clearance
Motorists' information

3.4 Ideas generated by the Workshops

Participants at the workshops were encouraged to contribute with their own ideas during group discussion about what measures might be introduced to help ameliorate and resolve some of the above issues. These fell broadly under three headings i.e. operational improvements, technological improvements and training. These are summarised under the same headings as above and classified according to type of measure. Shorter-term operational measures are identified in blue.

Detection and verification
Response
Site Management
Traffic Management
Incident Clearance
Motorists Information

View larger image

Figure 3

3. A Better Understanding of Incident Management

4. Review of Incident Management in the US

4.1 Introduction

At this point, it is worth describing the incident management experiences in the United States of America. A review of 'Best Practice' is identified as one of the key study objectives and it is clear that in terms of demonstrating an understanding of the issues raised in the workshops the US experience comes closest.

4.2 The case for more formal IM procedures

The case for incident management in the US has for some time been presented by a variety of State authorities. There is some form of incident management activity in most areas which typically involves each agency carrying out its own responsibilities, with primarily working-level and middle management administrative teams to provide co-ordination with the other agencies who are involved in their own aspects of managing incidents. Such a situation may however achieve less than the full potential benefit, and also leave open risks for failure within individual agencies and on a broad scale.

The IM (Incident Management) programme therefore seeks to promote the development of a framework for more formal multiagency traffic incident management programmes, with endorsement by, participation from, and co-ordination by senior management which includes all of the participating agencies. The formalising of IM procedures therefore involves a transition from reliance on co-operative relationships existing between responders and between middle management of agencies to an official recognition and endorsement of incident management as a core activity at a senior level. IM is then recognised as an overall initiative and purpose within and across agencies, significantly increasing the likelihood that its influence will be recognised when policies and other programmes, which may impact upon it, are discussed. When IM reaches programme status it becomes an integral part of the planning process within each agency, including planning of Information technology needs, or defining the types of vehicles required on patrol.

Similarly, such focus demonstrates a long-term commitment to incident management. Formalisation moves IM from "special programme" status, subject to budget availability of resources from outside to the mainstream where its existence is not questioned at each budget and staffing cycle. It has a clearly defined strategy, linked to the regional and state-wide strategies guiding such major areas as law enforcement and transportation. The strategy sets the direction for IM, supporting the policy and resource, which increase its impact and effectiveness.

This programme then also becomes a component in the budget process of each participating agency to obtain resources to implement and sustain the programme. This executive endorsement, as well as the infiltration of IM into every major aspect of each agency, provides a foundation that sustains IM from year to year across personnel changes and even political changes. Furthermore, it creates conditions where agencies support one another's requests for the resources necessary to carry out their respective incident management duties and for resources to expand programme scope and coverage.

The primary goals of incident management are to minimise the impact of incidents and reduce the probability of secondary incidents. Six measurable objectives of IM are:

The steps for implementing and sustaining a regional IM programme involve:

Phase 1 - Programme Concept
Phase 2 - Programme development
Phase 3 - Programme Maintenance and sustainability

These stages are identified in Figure 4.

View larger image

Figure 4

The organisations typically involved in IM are:

Results of effective IM

The following performance data has not been verified and the data underpinning such estimated savings has not been checked.

Safety benefits are clearly of equal importance but were not quantified. Benefits would accrue through lower exposure to risk and shorter periods of exposure for incident responders and victims, and fewer injuries to motorists resulting from the reduced number and severity of secondary incidents.

4.3 Lessons Learnt

The diagram in Figure 5 below is often used to describe the congestion costs/benefits associated with incident clearance. Accident benefits in terms of reduced exposure to risk are less well quantified.

The US examples point to the phased activity of incident management as defined previously i.e. detection, verification, response, site management, traffic management, clearance and recovery.

View larger image

Figure 5

Some of the lessons learnt according to the different organisations are described below:

Transportation Agencies

Traffic operations centres can function as information management centres in support of multi-agency operations.

Because route diversion has been proven to be an effective incident management tool, state and local transportation agencies should develop a combined strategy and implementation plan for co-ordinated arterial signal control during incidents.

Using changeable message signs to provide additional incident information to motorists, such as estimated travel times, improves the value of the information and motorist compliance.

Law Enforcement

Law enforcement and traffic management centre personnel must be co-ordinated, not simply collocated. In some places, patrol officers are rotated to work in the TMC (Transportation Management Centre) and to experience the technology firsthand. This means that they are more likely to radio the TMC for more detailed information on the location and severity of the incident.

Law enforcement resources will be available to respond to more urgent concerns by drawing on resources, such as close-circuit television for incident verification and service patrols to respond to disabled vehicles.

By closely co-ordinating with traffic management centre personnel after arriving at the scene of an incident, law enforcement personnel can improve on-scene command and control.

Use of law enforcement personnel on motorcycles for incidents occurring during peak travel periods can improve response times.

Service patrols

To be fully effective, service patrols must communicate and co-ordinate other activities with responding agencies, and should have access to the proper radio frequencies.

Outsourcing of patrol operations is beneficial because it:

  • Is easier to allocate and adjust resources according to needs
  • Minimises unit costs due to competition among providers.

Service patrols on bridges and in tunnels (where access is severely constrained) are critical to restoring the normal traffic flow.

The IM literature points to an impressive array of average response times, typically in the range 8 10 minutes. Again, these figures have not been checked.

Fire and Rescue

Fire and rescue are critical to the development of a cross-functional operations plan.

Fire and rescue agencies have considered modification of their operations procedures to better accommodate traffic management during incidents.

Aggressively seeking and maintaining the involvement of fire and rescue in multiagency planning and co-ordination will help ensure full co-operation in traffic incident management activities and programmes.

HAZMAT

Service patrol vehicles equipped with basic HAZMAT response equipment can more effectively manage the containment of minor spills and protect the incident scene.

Standard incident management procedures may need to be modified in order to accommodate the risks involved in working with and around HAZMAT incidents. Specialised techniques and specially qualified personnel are often necessary for safe and effective HAZMAT incident management.

Providing incentives for HAZMAT contractors based on timeliness and efficiency of speed of response and cleanup can minimise costs while maintaining performance.

Towing and Recovery

Because they are critical to rapidly restoring normal traffic flow, towing operators should be involved in interagency incident management training.

A hybrid of traditional and performance based contracting that requires operators to meet specific requirements (e.g. number of vehicles, response and times, storage space, insurance and licensing) can improve responsiveness and reduce cost.

4. Review of Incident Management in the US

5. Main Analysis

5.1 Introduction

This section focuses specifically on the tasks involved in 'Incident Management'. The preceding section raised a number of issues which came directly from the workshops in terms of roles and responsibilities and the need to co-ordinate better the planning of incident management. A key step in this process must therefore be to have a much clearer understanding as to what actually goes on 'on-the-ground' during the unfolding of an incident.

The data set is large and has involved representation and analysis of data from a range of sources. The principal findings are summarised by the analysis/analyses from which they are drawn. Naturally this leads to a certain amount of repetition which should be seen as increasing the weight of evidence for a particular viewpoint.

5.2 Main Task Analysis

Defining the task of incident management

As stated earlier, defining an 'incident' is not a simple process. This is an important finding in itself. Top-down analysis can lead to the adoption of criteria such as 'any event to which at least one emergency service vehicle is summoned', or 'any event that has a negative impact on traffic flow that needs some form of positive rectification'. There are many variations but these two examples illustrate a tendency to define the event in terms either of the generic resources that might be required to deal with an incident or its impact upon the performance of the road traffic system. As such they bracket the key area of what it is that actually happens in an incident and the process of managing it. There is nothing wrong with this type of analysis. It leads to an extensive and very valuable consideration of what might be the best ways to deal with an incident. This is readily exemplified by the US Traffic Management Incident Handbook which is written by highly competent and experienced professionals and which clearly has an insight into the way in which incidents unfold, how they impact on the road traffic system and the range of interventions that might be needed to improve incident management performance. The content of this handbook in many ways reflects the UK experience in terms of both recognising the systemic importance of many issues, for example training and inter agency co-operation and the complexity of organisational issues. Some of the detail is very reassuringly congruent with information found elsewhere in this report.

Notwithstanding these positive aspects, there is a problem with the approach of defining incidents by reference to resources required to deal with them and by their impact of system performance. The first problem is that no attention is paid to what actually occurs when an incident is managed. This is because it is quite variable and as noted elsewhere incidents are not homogenous. This is compounded by the fact that good intentions, as expressed by procedures, quality standards and the like are not always adhered to for a whole range of reasons that include task familiarity, illusions of invulnerability, the impracticability or appropriateness of such standards and the impossibility of following complex procedures in high workload, stressful situations. Consequently, it is easy to create organisational and procedural edifices that give the illusion that the problem is being dealt with when the reality is that people who undertake incident management work must work around the procedures. We remain relatively uninformed of the actual tasks that comprise incident management. The second problem is that while interagency co-operation is identified as a key issue, the notion of how planning and conducting operations occur both before and during an incident is not fully understood. It is quite clear that planning incident management can take on board idealised views of cross-organisational co-operation and interaction that does not map very well onto the reality of the situation. The analyses carried out in this study collect data from a variety of sources that directly access the experiences of those who undertake incident management as well as those responsible for managing incident management systems and planning implementation. There is a marked contrast between the results of a bottom-up task analysis and that of a top-down systems viewpoint. This disparity is a key finding as it opens the door for a more eclectic analysis that includes the actual details of what occurs in an incident as well as the planning and system performance aspects. This will facilitate the consideration of more directed interventions and methodologies. The detailed task analysis reveals a simpler set of task features than might otherwise be expected that should allow incident management to be improved significantly in an incremental manner customised to local needs.

Opportunity to make impact after an incident begins

The material and human resources that are available at the start of the incident place effective limits on the speed and effectiveness with which incidents are handled. Once an incident is underway, off-site management based interventions have little potential for impact upon the management of the incident itself. This may sound obvious. Its importance lies in highlighting the responsibility of non-operational management to identify the elements needed to ensure good incident management performance are in place. This is a common failing in the management of safety critical systems and normally results in unrealistic demands being made of operators, be they pilots, industrial operatives or in this case traffic police. As an example consider a traffic sergeant trying to organise the arrival of capital equipment through a series of control rooms and civilian 'on call' personnel using a combination of two mobile phones and his radio. This adds workload, complexity and increases the opportunity for misunderstanding and delay. The responsibility for failure will fall on the shoulders of the traffic sergeant.

Phased activities

The core incident management tasks considered here, defined as being those occurring once the 'incident manager' (police officer) reaches the incident, can be broken down into four distinct phases:

The bulk of the tasks occur in the high workload Phases A, B and C. The quality of performance in these three phases, which may all occur during the very early part of the incident, can have a major impact upon downstream clearance times.

Differing nature of roles

There are two types of role in incident clearance. The more numerous specialist roles filled by FRS, ambulance, recovery, road maintenance and that of the de facto incident manager. This latter role is normally filled by a police officer who takes overall responsibility for the situation. The specialist roles are convergent and involve goal-oriented, highly focused work with a clear assessment - action - completion format where good performance is easy to assess. The 'incident manager' role is divergent and involves creating an ongoing, transient organisational infrastructure in which the other specialists (which may well include police in other roles) undertake their work. In one observed incident, the FRS was releasing casualties and making vehicles safe; the ambulance service was assessing and stabilising those who might have injuries; some police were taking details and witness statements where this proved to be possible; other police were managing the public, the incident ground, the traffic flow on the other side of the carriageway, under the supervision of a traffic sergeant who was repeatedly checking the status of events with colleagues and with control and planning for future need. His work was overseen by an inspector. The incident management role contains conventional management tasks such as delegation, checking, and monitoring, giving feedback, ensuing that plans of one kind or another have been executed and facilitating timely and appropriate communications to all parties including the control room. It also includes problem solving.

Mental models of events

The key cognitive skill of the incident manager is to build and maintain an accurate mental model of the incident so that appropriate situational knowledge and knowledge from other domains and constraints can be integrated into the decision making and planning process. Other key cognitive skills are the ability to scavenge for information repeatedly and proactively; to maintain and test the veracity of the models; the ability to cope with and reduce areas of uncertainty; the ability to deal with multiple streams of information; the ability to prioritise and the ability to construct rapidly and modify short and medium term action plans on an ongoing basis. This may be a simple task in small incidents but in large incidents, it is a considerable mental activity and one that is not wholly apparent to the observer but none the less critical for that. The occasional suggestion that police officers stand around doing nothing while the other emergency services do their work is inaccurate.

Formal characterisation of the incident managers task

Figure 6 provides a representation of the task of the incident management officer. The core of the task of the incident management officer is an iterative process of assessment, checking, reviewing, planning, communications and implementation of actions. This iterative cycle influences the mental model and as time progresses, draws upon the mental model. The left-hand column is a range of action strategies available to the incident manager that will vary from one incident to another.

View larger image

Figure 6

Potential for error

Table 1 allows different types of 'error' to be unpacked. It seems likely that the bulk of the errors will result from misconceptions and misunderstandings or failure to communicate appropriate instructions to the team. This is an important distinction as the defences against cognitive and conceptual errors lie in the design of training, equipment, procedures and organisational communications, all of which are things that need to be in place prior to an event happening. This reinforces a key finding that once an incident begins to unfold many of the principal determinants of how successfully it is handled are already in place.

Table 1: Analysis of Errors
Cognitive / Conceptual Errors Communications Errors Task Based Errors
Misunderstandings of situation Incorrect information Errors of execution
Misunderstandings of intentions of others Misperceptions Errors of commission
Missing key decision points Ambiguity Errors of omission
Mistaking past intentions for past actions High workload  
Reacting inappropriately to complexity Competing information  
Bad prioritisation    
Stalled planning and decision making    
Confirming, not testing, an initial hypothesis    
Being anchored to an initial interpretation of events    
Being overly influenced by partially similar past    
experiences    
Cognitive oversimplification of complex situations    
Multi-personas in incident management role

Responsibility for the management of incidents is distributed over time, over different people and different organisations. Although traffic police normally arrive first on motorways and trunk roads, in rural areas the FRS (possibly retained) may be first to the scene as may be section officers in urban areas, especially if traffic officers are engaged on other duties. The control room function also has some form of incident management responsibility. Control of complex incidents will be handed over to a senior sergeant sooner rather than later. In addition a more senior officer may come out to large incidents. Functional liaison occurs at task level, supervisory level and management level. The mantle of control is thus both distributed and mobile. The context of each incident is different and so it is not possible to specify how things should ideally occur, however tempting this may be to do. This is one of the problems of defining incident management from the viewpoint of resources that might be deployed when compared to an analysis of what actually goes on in the management of an incident. It is possible to end up with a hierarchical set of tasks that might be undertaken under ideal circumstances and not address how in fact they might actually be dealt with.

Impact of need for formal accident investigation

The need to collect evidence for a later court hearing will add to any clearance time. It seems likely that formal accident investigation (AI) is not a frequent occurrence. Detailed area searches if needed are obviously time consuming, although they are not common. The normal activities of an AI officer can be undertaken quite quickly. They naturally include a topographical map of the area of the accidents and of relevant surrounding areas. This can be achieved rapidly with sketch book, measuring tape and camera. The hi-tech notion of some form of scanning device to map the area may be both expensive and unnecessary. One of the major problems seems to be getting the AI officer on site in a timely manner, as such individuals are apparently rare. This can lead to delays in attending while other tasks are completed or while they are rostered up from 'on call status'. Once on site they proceed quite quickly, so early arrival is a key. Other tasks necessary for the investigation include traction tests, examination of road surface and other diagnostic activities. Coordination and organisation of taking witness statements and ensuring that all witnesses are contactable is also a key part of the AI activity that will require police resources. At the present time AI officers are police officers.

Organisational heterogeneity of response team

This might apply to any incident but it is a major issue in larger events. This is because more agencies are needed to deal with a large complex incident with increased numbers of specialists being involved. It is not just a question of mobilising resources, as would be the case if it were the responsibility of one large agency to deal with road traffic incidents. There are organisational boundaries to be negotiated. A simple example concerns the rotation list for recovery vehicles where clearance work is awarded on a strictly rotational basis that takes no account of the particular competence of the vehicle clearance contractor or of the location in relation to the incident or the resources that are currently deployed. Vehicle clearance contractors have targets (for example to respond within 30 minutes). These appear to be applied quite flexibly and at the discretion of the officer attending the scene. It is therefore possible for a contractor to take rather longer, but for the delay not to be logged, in the interest of maintaining harmonious working relationships. The situation thus arises where a performance target appears to have been met when it has not. In the type of minor incident where this might occur, the impact upon congestion may almost double. More complex examples arise when specialist equipment, road clearance resources, repairs to safety equipment, and so on are needed for clearance. Poor planning and foresight may give the impression that organisational heterogeneity may contribute to delays. It seems likely that if the problem is recognised and the question formally asked 'how can we deal with these issues' then any difficulties are amenable to interventions. Such interventions are unlikely to be cost neutral.

Platinum twenty minutes

Much is made of the notion of a golden hour during which the middle group from a victim triage might survive. It seems likely that there may be a platinum period, perhaps 20 minutes, perhaps 30 minutes at the beginning of any incident when properly executed incident management may mitigate and minimise effects on down stream traffic congestion. The key to the notion is to reach the incident, carry out primary, secondary and tertiary assessment, acting as needed within the first 20 or 30 minutes (see example timeline of incident management tasks in Figure 7). This can apply to both smaller and large incidents and as a performance target makes a great deal more sense than trying to define an overall clearance time. It should prove possible to specify the key tasks that a Platinum phase of incident management should contain. Generally speaking the end of the platinum phase will be when the incident moves into a routine, wholly managed status. Thus generic tasks and training for the Platinum period should prove readily specifiable. Good platinum performance will depend on the organisation structures, ethos and relationships of the key stakeholders. Good platinum performance will probably predict good clearance performance. The interventions that will lead to good platinum performance will integrate training, development organisational culture cross-organisational relationships, functional management and communications effectiveness.

View larger image

Figure 7
 

5.3 Workload Analysis

The results of the workload analysis are quite straightforward and not unexpected. At the beginning of an incident, there are a whole range of predictable routine tasks that need to be performed by a range of key stakeholders. The key parties are the control rooms, the police and the other emergency services. The variable nature of each task means that while the activities are generally predictable, the exact actions required will vary from one situation to another as will the order in which they are undertaken. Workload is very high and experience, training and pre-incident competence combine to ensure prompt and timely responses to complex situations.

Awareness of clearance needs figure in the tasks from an early stage. In smaller incidents where there is no injury, minor damage and only the 'normal' risk of being abroad on a fast moving highway, clearance may be the first task that is considered. Even on major incidents, clearance as a serious issue becomes a major question from very early on, perhaps 15-20 minutes. Any suggestion that it is a last minute activity is inaccurate. There is a serious issue here. It lies in the correct early assessment of clearance needs, and so of resources that must be marshalled, and the traffic sergeant or other incident officer trying with the aid of control to reach and co-ordinate a range of different resources. Although this should be a control based task, it does become a major and time-consuming concern of the incident manger officer who may check back constantly to make sure things are moving. One of the difficulties is in reaching a diverse range of different organisations. This is discussed elsewhere in the report, but it makes major contributions to the workload. It clearly is a matter of some concern to the officers concerned. They also report that they have no desire to remain in a potentially dangerous situation any longer than necessary and are aware that they will be held responsible if there are extensive clearance delays.

High workload in itself does not seem to cause or lead to delays as long as the incident comes under the control of experienced incident management officers at an early stage. Inexperience can lead to inappropriate actions that make secondary incidents more likely or actions that have later consequences for example not gritting the road backwards from a diesel fuel spillage. After the initial burst of high workload, that maps onto phases A, B and C and may last anything from 5 to 60 minutes, tasks become routine and constrained by factors other than immediate required actions. At this point if all resources needed were immediately available the road could be reopened very quickly. Workload is not an issue. Quite the reverse. Police officers have no desire to wait around doing nothing while they wait for other agencies to arrive, more so as they know that motorists who pass will wonder what they are doing and assume they are wasting time. It is not unknown for reports from traffic planes to make comments about 'it looks clear why don't they open the road', or for VIPs (through special channels) to complain about the apparent lack of action after they have been caught in a delay.

To reiterate, workload is not a problem as long as reasonable resources and experienced incident management professionals are managing the incident. High workload might severely damage both safety and clearance times if those managing the incidents in its early stages lack experience. The first 20 Minutes is the key. As noted elsewhere there may be a 'Golden' hour for casualties but there is a 'Platinum' twenty minutes for incident management.

5.4 Task Categorisation Analysis

The task categorisation analysis examined the individual task outputs from the main task analysis and categorised them into a range of different types of task. This used criteria derived from cognitive psychology and the expert judgement required to analyse the tasks and sub tasks that make up the elements of the overall task of running any safety critical system. The tasks were analysed two levels down in the analysis, not three for a number of reasons.

Firstly, not all tasks could be decomposed three levels down, and the inclusion of some tasks at a third level of decomposition (i.e. more detail than the second level) would have upset the emerging taxonomy of task classification by over representing very fine tasks. Secondly, the level of decomposition achieved on the third level analysis is perhaps too detailed to be is used in this type of analysis. The methodology involves a process of taking each individual task in the listing from the main task analysis and describing it in a parsimonious and generic manner, avoiding over elaboration and context specific characteristics. For example directing members of the public to move to a certain location might be described as an 'instruction' to others to 'act'. Monitoring the flow of traffic through a choke point might be described as 'cognitive monitoring' task. Chasing control rooms to confirm that others might have taken certain actions might be described as 'checking plans'.

As the number of tasks that are classified grow, patterns emerge that allow different small categories to be collapsed into one another and by an iterative process of refinement it is possible to take the output of the task analysis and end up with a simple summary of the characteristics of the tasks. The direction in which the analysis evolves depends on the goals of the analysis. Naturally different individuals produce different details within their individual attempts at analysis, but these converge on a common themes as the analysis progresses. After a number of iterations, the analysis converged in three types of category. These were:

(a) Information processing tasks.

Information processing in this context refers not to information processing in the cognitive or mental sense but in the organisational sense, in other words passing and processing information around the system, be it giving instructions, advising on completion of a particular task, receiving information and so on.

(b) Action Management tasks

Action management tasks involve all the normal management functions, such as planning, monitoring and checking, evaluating, prioritising, planned allocation of priorities, coping with undetected circumstances, forming a conceptual model of what is going on and so on. All these tasks are cognitive or mental and what characterises them is the tendency for them to be unobservable by an outsider. They may well be carried out without the full awareness of the skilled personnel who are undertaking the work. They interface with Information Processing tasks and with Actions tasks.

(c) Actions tasks

These involving actions or causing action to be taken and may be expected to result from action planning and communications inputs.

The results of the analysis of task and subtasks by these categories are given in Table 2 below under the main task headings. The results of this analysis are very informative. Reflecting the main task analysis, the tasks are mainly though not exclusively undertaken by traffic police in the data samples accessed in this study. Also Ambulance Service Tasks are not included because apart for ensuring access they have a very self-contained task that is not part of the incident management task. In contrast the fire and rescue services do tend to be more involved in incident management.

Table 2: Results of Task Analysis
Main Task Information Processing Tasks Action Management Tasks Actions Total
Awareness of incident 16 16 1 33
Location of Incident 9 18 3 30
Mobilisation of Emergency Services 3 6 3 12
Guidance of Emergency Services 1 6 2 9
Getting to incident 0 6 4 10
Operational liaison 3 1 1 5
Make initial assessment at scene 3 6 1 10
Create safe work environment 0 7 4 11
Make secondary assessment 0 7 0 7
Act on secondary assessment 2 2 6 10
Manage traffic flow 0 18 1 19
Make tertiary assessment 0 5 0 5
Act on tertiary assessment 1 4 0 5
Manage initial arrive of FRS 0 3 3 6
Manage members of public 0 2 5 7
Construct infrastructure 0 17 2 19
Make quaternary assessment 0 10 0 10
Act on quaternary assessment 3 1 0 4
Main phase - Police 0 13 0 13
Main phase FRS 5 3 4 2
Accident Investigation 2 5 18 25
Clearance 4 14 6 24
Inspection 0 0 5 5
Declare road open 1 4 0 5
Open road 0 3 2 5
Manage distal effects 0 2 1 3
TOTALS 53 179 72 304

The results speak for themselves. Of 304 individual subtasks for the composite main task analysis, 179 (59%) were Action Management Tasks, 53 (17%) were information processing tasks and 72 (24%) Action Tasks. Having conducted the analysis this seems unsurprising but it is a finding that may radically change the way in which we view the nature of incident management and so how we might seek to improve it. This is because it is quite clear from the limited information that is available from both the US and the UK on preparation for incident management that all analyses tend to concentrate on command structures and the completion of the concrete tasks that experienced individuals know underlie incident management. The information processing tasks will be lumped together under a general and unexpanded heading of communications and the management tasks that are not visible to the naked eye but that make up the bulk of the tasks, are largely undiscussed. This analysis explains an observed disparity between traffic officers, who are undertaking the work, and the world of operational analysis where the analyses conducted fail to give due importance to the cognitive skills and operational competencies of the traffic officers. The traffic officers are left with the correct perception that they are undertaking a skilled and specialist task that seems unmentioned and acknowledged by the analysts. This gap maps onto the cognitive aspects of the task which as we can see form a significant portion of what the person managing the incident actually does.

A number of conclusions follow from this revealing analysis. First, if we are to improve incident management performance we have to make sure that those who are managing the incident are properly trained in undertaking cognitive, action management types of activity, including not only basic cognitive competencies, but also an awareness of errors, faults, risks and how they may occur and how they can be managed, so they can be best avoided. This element of the task is virtually un-addressed in any formal training regime that we have located and is learned by 'sitting with Nelly', on the job. As such it is unrecognised and unvalued except amongst the actual practitioners. It is accepted that not everyone who becomes a traffic policeman actually stays the course. It also follows that putting people without proper experience and training into incident management situations which has been mooted in a number of forums, will not only be risky in terms of their own well being but also that of others involved in the management of incidents. Finally, if we wish to improve incident clearance, a significant part of incident management activity, it is unlikely be improved if we do not explicitly address 59% of the task, pay minimal attention to 17% of the task and only attend in detail to 24% of the task which in any case depends quite extensively on the other two components. Prior training and organisational systems development are the two established methods for addressing these issues in the safety critical areas of aviation and power generation:

  1. A full task analysis of the information processing needs of the incident managing task, including all stakeholders organisation will allow improvements to be made;
  2. Training and developing cognitive skills and error reduction amelioration and minimisation and;
  3. Rationalisation of existing skills/competence provision for those responsible for managing incidents. This is essential.

5.5 Solutions Task Analysis

The analysis of the different types of intervention derives from the overall task analysis and simply takes each of the top level task elements that are listed in the task analysis and asks the question about the type of intervention that might be designed and implemented to improve both the management of incidents generally and of reducing clearance times in particular. The two are of course wholly congruent. The findings may be summarised as follows.

On this analysis the opportunities for technical innovation are relatively few and are confined to small aids rather than major root and branch, technically led, changes to the entire systems. This is due to there being relatively few major technical innovations that might be implemented and the fact that the study has examined in some detail the way in which incident management is currently conducted. The types of technical intervention that might be useful involve interventions concerning the better use of the systems for advising drivers and motoring advice organisations about delays. Note that excellent technology is wholly dependent on the timeliness, accuracy and appropriateness of the advice that is received and so its use is intimately related to organisational matters. For example the benefits of extra traffic cameras may be smaller than would otherwise be the case if the information is not passed on promptly and modified as it changes. Similarly automating the system with speed restriction and posting advice to motorists that relies to a certain extent on remote sensing is very hard to implement as it needs a vast surrounding infrastructure that probably does not exist in any location in the UK. Furthermore, the design and implementation of the required semi-automatic algorithms may well end up causing more problems than they resolve due to unforeseen aspects of system performance. This would not be the first time such a thing had occurred. Technical innovation and automation can be very useful but their implementation can be capital intensive and only partially effective, sometimes performing no better than a previous system.

Training and the development of skills and competencies seem to be the major opportunity for developing good levels of performance. Training, though presently well intentioned is piecemeal across police and other parties. It concentrates on actions, rather than planning and understanding of incident management. It appears that most of the effective training occurs 'on the job' as less experienced officers are informally inducted into the competencies of traffic work and incident management by a group of more experienced colleagues. This occurs by supervision from more senior colleagues, from rest room and other exchanges and of course from working with a more experienced officer. As noted elsewhere this also acts as a filter for those who are not suited to traffic work that return or move on to other duties after a few months. This is quite effective but is wasteful of resources if it is indeed a major issue. The key to successful training lies in addressing the key parts of the traffic management task which are principally concerned with detailed aspects of planning and monitoring and other aspects of management that are cognitive in nature. Competence in actions seem to be relatively easy to define and undertake. It seems that physical tasks (e.g. 'cone this way or that', 'lay out protective vehicles in such a manner', 'delineate a sterile zone of certain dimensions' and so on) are the main focus of such training and this might be usefully developed.

Organisational issues (of which training is really a subcategory) is the other significant area where incident management performance might be supported and improved. The general areas where incident management can be designed include the development of team-based working in a number of dimensions; appropriate and timely supervision and support from colleagues to help preview situations or provide further information; the definition and optimisation of working relationships between different organisations and timely organisational responses to events. Other specific organisational issues include ensuring that all colleagues are properly trained and experienced to do their work; that they have the basic aptitudes needed to do the work; appropriate aids in the form of checklists, guidelines, procedures and appropriate decision making aids are in place. Provision should be made to deal with peaks of workload, although this can be a difficult problem to solve though it is quite possible to do so. Intra- and inter organisational boundaries should be delineated and mapped and a local analysis made of the potential difficulties that might occur and provision put in place to deal with them. Once again planning is the issue. Just because a system is set up does not mean that it is suitable for the task for which it has been designed. Testing and bedding down organisational systems is an ongoing activity. An example would be looking at the difficulties an officer managing an incident might have in co-ordinating recovery, cleaning and highway repair service from the incident. The key is a process of organisational review, asking those with approximate knowledge and experience what the problems are and then designing ways off actually dealing with the problems.

5.6 Top Level Task Analysis

Error analysis summary

The top level 'main' task analysis highlight two types of information. First the types of failure/errors that might occur and second, the types of impact on clearance times that various failures might induce. With regard to errors, things that might go wrong, the key types of error that might occur are as follows:

  1. Errors of planning are far more likely than errors of execution. In other words things that go wrong tend to be complex and the result of misunderstandings or misconceptions or unforeseen events, impacting on otherwise suitable plans.
  2. Errors of omission are likely to be more common that errors of commission. In other words, colleagues forget to do something rather than do the wrong thing. One variant of this is thinking you have done something when you have not. These types of errors tend to be rectified as time passes and arise because of situational complexity and workload peaks. They cause simple delays and compound delays, the latter being additional time lost caused by failing to deal with a situation rather than a simple delay where a five minute delay in undertaking a task leads to a matching five minute delay down stream.
  3. Misunderstandings and misconceptions of the situation may lead to inappropriate actions and delays and it is clear that one of the key parts of the task of those responsible for incident management is maintaining and updating a valid mental model of the incident as it unfolds. This requires proactive checking and other cognitive skills.
  4. It seems likely that people with appropriate experience cope well with the high workloads of early incident management. Novices may find it rather more difficult as the situation unfolds.
Impact on clearance times.

Impact of clearance times is both a simple and complex issue. Each situation is unique with its own context but some patterns do emerge. First the sooner a skilled and experienced incident manager arrives at an incident the less likely there are to be delays. Less skilled colleagues may be slower and make 'poor' decisions. Small delays in the early stages can have major amplificatory effects on the congestive impact of the incident. It is tempting to make a distinction between health and safety and clearance needs and as noted elsewhere in this report, clearance needs are considered quite early even in complex incidents, but they are considered as a growing priority, that start to make an appearance after about 30 minutes. In fact given the importance of secondary incidents and the harm and injury that can occur with such incidents clearance is an important issue with regards to health and safety, though obviously they will be superseded by the immediate needs of an accident victim. Generally time is the key to dealing with injury, safety and clearance. Not so much a simple response time metric that can be damaging, but the notion that getting a certain set of tasks undertaken within a certain temporal envelope may be more useful. The key to dealing with incidents (unsurprisingly) is to have the quality systems, organisational interfaces, human resource and training in place and available to respond before the incident occurs. How things are at time zero in an incident largely determines what will happen. Thus incident management is really an organisational issue. However little evidence of sophisticated analysis for organisational prerequisites were found. It is in this area that the greatest gains can probably be made, using something like the suggested Platinum 20 minutes as a focus for this.

5. Main Analysis

6. Conclusions

The remit of this study was to examine what it is that actually goes on at a road traffic incident, with particular reference to the clearance of the incident and restoration of normal road conditions, at incidents where one or more emergency vehicles attends. The question of target setting for clearance time was one of the issues driving the investigation, as it had been suggested as a mechanism for somehow galvanizing clearance performance and motivating those responsible for incident management to clear incidents rather more quickly than would otherwise be the case. This it was felt would bring economic benefits.

A number of clear specific conclusions can be drawn from the data gathering and subsequent analysis and these are as follows:

  1. The incident management role is distributed between different parties, mainly control rooms and one or perhaps more individuals at the site of an incident. De facto this often devolves ultimately to a traffic police officer on site. Police officer and control room teams offsite also hold responsibility and the boundaries between on site and off site are somewhat fluid. This fluidity can add to the workload of the on site officers who feel the need to chase and check progress, as they are the ones who will be seen as holding responsibility in the case of any difficulties.
  2. Because of the need to exercise police powers in managing incidents, it is difficult to see how a 'yellow light' contractor can manage a significant incident without the aid of the police, unless they are granted some form of statutory authority to act on highways, for example the authority to stop traffic, direct members of the public and to co-opt resources. Some form of dual management might prove useful however.
  3. Incidents are complex and while they have common features, each one has its own complex context. Over simplified generic solutions will hinder those responsible for managing incidents rather than help them. Examples of such solutions include simplistic target times for clearance and the use of high technology aids on which a proper User Requirements Analysis/Specification has not been conducted. While a start point for an incident can be specified, it is very hard to say just when the incident finishes.
  4. Interagency co-operation, normally involving the three emergency services, ordinarily proceeds effectively in the early stages of the incident especially when there is a risk to health and safety and victims require attention. Unfolding organisational complexities may though lead to delays.
  5. The role of the traffic police, normally responsible for incident management, is a divergent role that involves the creation and management of a transient infrastructure in which the incident is managed and activities undertaken. Based upon figures in this analysis 58% of their tasks are cognitive tasks that involve not action but review, checking, planning and integration. 24% are action tasks. The remainder are communications type tasks. Most of the training and development for incident management revolves around the action tasks and to a lesser extent the communications tasks, leaving the cognitive management tasks unrecognised and undervalued.
  6. Other emergency services have a relatively small impact on incident clearance. The ambulance service generally arrive promptly, do their work rapidly and depart, for obvious reasons. Fire and rescue services present promptly and conservatively leading to some difficulties. These include occasionally being first at an incident and over protecting the scene so causing excessive congestion, responding with multiple appliances that then cause delay by virtue of their bulk, their prominence and their poor manoeuvrability when reaching, attending or egressing the scene of an incident.
  7. The early stages of an incident are very high workload. Many key tasks have to be performed by control room operators and traffic police, and there are a number of phases that may unfold in the first twenty mutes. These have been detailed ea