Highways Agency

Chapter 3 - Assessing Risks

3.1 This chapter considers the risks and consequences associated with nearside accidents and different methods of assessing risk. Having considered the existing standard and an analysis of accident data (see Chapter 2), the Group needed to consider the risk and probability associated with nearside accidents, to inform their considerations about the standard and any need for change. An important part of this consideration are the consequences arising from accidents and their effect on others (third parties), especially rail operators.

Risk and Provision

3.2 A risk is the chance, great or small, that someone will be harmed by a hazard. In the context of the standard, a roadside hazard is any physical obstruction that may, in the event of an errant vehicle leaving the carriageway, result in an accident causing injury to the vehicle occupants. In the case of nearside accidents, the roadside hazards are features along the side of a road, such as lighting columns/lamp posts, bridge supports, telegraph poles, trees, road signs, ditches or other water features, embankments or cuttings, the ends of bridge parapets and any other permanent object. As has already been discussed in Chapter 2 and is shown below, safety barriers on the nearside of roads reduce the chance of an interaction with a hazard but also in themselves carry some risk of injury to the occupants of vehicles striking them. The Group therefore had to consider both the probability or frequency of nearside accidents and the scale of the resulting consequences.

3.3 The estimated total number of vehicles leaving the nearside of English major roads in any one year is about 3400 (see Annex 5 Table A5.2). This figure is based on STATS 19 accident numbers that have been revised upwards to take account of the under-reporting of accidents (see Annex 5 Section 2 for the calculation of the adjustment). The estimated number of single vehicles leaving the nearside of major roads and striking a safety barrier is about 500 vehicles per year, about 15% of the total, while the number likely not to strike a barrier is about 2900 (see Annex 5 Table A5.2). The 15% is reasonably consistent with TRL's observation from a road survey of sections of the M3, M4, A339, and A34 that, despite a wide variation between dual and single carriageways, about 17% of the major road nearside carriageway in England is currently provided with barriers (see Annex 4 Section 1). The estimated number of vehicles striking safety barriers is roughly in proportion to the estimated amount of safety barrier on English major roads.

3.4 The Group were interested in the probability of an accident where an errant vehicle reaches and blocks a railway line. The probability is made up of the combination of a number of single events occurring sequentially. Hence, the probability of a vehicle reaching a railway line would be the combined probability of a vehicle leaving a major road on the nearside, not striking a roadside hazard or safety barrier and travelling a distance to reach a railway line. Figure 3.1 shows a comparison of the estimated number of vehicles per year leaving the nearside of major roads in England, hitting a roadside hazard, hitting nothing, hitting a safety barrier, or reaching a railway line.

Figure 3.1 - Comparison of the Estimated Number of Single Vehicles Leaving the Nearside of Major Roads and being involved in a Personal Injury Accident (PIA) taken from Adjusted STATS 19 Data (see Annex 5 Table A5.2)

The analysis in Annex 5 section 3.5 shows that the probability of a road/rail accident of this nature is 0.4 per year. This is equivalent to approximately one vehicle every three years reaching a railway line, from the nearside of an English major road. This does not consider the probability of the vehicle being struck by a train or causing a derailment (as this event is being considered by the HSC Working Group). The Group considered that a vehicle leaving the nearside of a major road and reaching a railway line was an extremely infrequent event, relative to the likelihood of other types of road accidents on major roads.

3.5 The Group were also interested in any differences in the severity of accidents involving safety barriers compared to other roadside hazards. The number of accidents and injuries is recorded by severity in STATS 19 for all accidents involving personal injuries. A measure of the severity of an accident can be estimated by calculating the severity index, which is the number of fatal and serious injuries as a percentage of the total injuries recorded. Assuming the proportion of accidents is of the same order as the recorded number of injuries, then the proportion of single vehicles likely to leave the nearside of an English major road, not strike a barrier and result in a fatal or serious injury accident is about 21% (see Annex 4 Table A4.4.17). The proportion likely to strike a safety barrier but still result in a fatal or serious injury accident is about 18% (see Annex 5).

3.6 The relatively small difference between the severity of injuries arising from accidents not involving nearside safety barriers and accidents involving nearside safety barriers should not be interpreted as meaning the provision of nearside safety barriers is not worthwhile. Since safety barriers are generally provided at sites where the risks are greatest, it is reasonable to expect that without them there would be an increase in the number of serious injuries for the 82% of vehicles that are currently likely to hit barriers without resulting in a fatal or serious injury accident. However, the figures do suggest that providing nearside safety barriers along all major roads would produce only a marginal benefit, if any, in terms of preventing injuries or reducing the severity of injuries. The Group considered that this analysis supported their views reported in Chapter 2 and suggested that the existing policy on major roads of providing safety barriers to protect against particular hazards strikes the right balance (see also paragraph 2.33). They acknowledged that specific hazards of exceptionally high consequences (for example bridges on major roads over railways) may need to be treated as special cases. In such instances a risk assessment may be required to identify the locations or roadside hazards that may need particular attention.

Risk Assessment

3.7 Risk assessment requires possible events to be broken down into logical groups, for each of which the risks can be identified. The possible hazards are examined and the associated risks (which may relate to safety, time, costs and so on) are evaluated. A risk assessment at a specific site should identify and record what the problem is, the potential risks at the site and whether the risks have been considered and eliminated or, as far as is practical, mitigated. If that risk assessment forms part of a programme of assessments that is identifying mitigating work, then the assessment should also provide a means of prioritising that work.

3.8 The Group were unable to identify any explicit risk assessment in use for determining the provision of nearside safety barriers on major roads in the UK. The existing standard establishes a set of criteria under which safety barriers should be provided. In effect, a risk assessment has already taken place during the development and establishment of the standard, as the priorities for hazard protection in it are based on research and analyses of accident data. The standard gives minimum requirements for the provision of barriers, however there will be cases where additional provision is required. The designers are expected to use their experience in identifying locations where more than the minimum requirements are to be provided. Site-specific risk assessment could be used to determine the level of provision in such cases but at present there is no recognised methodology. Although the Group found no evidence that the standard as such had led to unsafe practices, they considered that a formal risk assessment method for the provision of safety barriers would be helpful to all those involved in making decisions about the provision of nearside safety barriers.

3.9 The Group were also unable to identify any explicit risk assessment in use to address all the risks at sites where road bridges cross railways. The UK highway authorities do currently use risk assessment to prioritise programmes of upgrading bridge parapets across their networks. The assessment model looks at the potential hazards associated with a site and then applies a predetermined expected frequency of occurrence and consequences. In broad terms it attempts to predict how hazardous a site is.

3.10 The rail industry uses risk assessment to help make safety decisions, based on historic accident information. There is a computer-based Safety Risk Model that provides the causes and consequences of potential accidents and shows how risk can be reduced, in terms of the costs of avoiding fatalities and injuries on the railway. The Group considered that a method of risk assessment that is common to both the road and rail industry, that would identify and prioritise the risks at sites where there is the potential for conflict between major roads and railways would be helpful to all parties. The Group acknowledged that, for a method to be acceptable to all parties, it would have to include a means of comparison of these risks with other risks on the road and rail networks.

3.11 The Group recognised that a risk assessment approach for minor local authority roads would also be beneficial, although it was not part of the remit of the Group. A methodology that is currently being developed for local authority roads was briefly reviewed. The Group discussed the principles of this draft framework and suggested that the DTLR should ensure that necessary further work by the appropriate bodies be undertaken to produce a risk assessment methodology that could be implemented for local authority roads. It would be useful if this assessment methodology could be compatible with risk assessment methodologies for major roads and the Group considered that this may provide a starting point for work on major roads.

Consequences

3.12 When an accident occurs there can be a range of consequences, related to the severity of the accident and other factors such as who and what is affected. The Group were interested in comparisons between the consequences of accidents involving roadside hazards and accidents involving third party property or equipment (such as the railways), which might be expected to be further away from a road than identified roadside hazards. The Group considered a comparison of the economic value of risk for various theoretical nearside accident scenarios on English major roads. The scenarios could be described as:

• Roadside Hazard - a "typical" nearside accident involving a roadside hazard, with few fatalities or serious injuries,
• Off Road Hazard 1 - a nearside accident, involving a third party off the road (say a house or a shop), with few fatalities or serious injuries,
• Off Road Hazard 2 - a major nearside accident involving a third party off the road (say a crowded bus stop, school playground or car park), with significant numbers of fatalities and serious injuries, and
• Road/Rail Hazard - a major road/rail accident, with significant numbers of fatalities and serious injuries and with considerable financial and social losses.

Annex 5 demonstrates how the economic risk value of each scenario was calculated as the product of the probability and the estimated value of the consequence. The scenarios used similar injury numbers, unit rates and estimated total consequential costs of disruption and other losses of typical events to produce an economic value of the risks.

3.13 The analysis suggests that in numerical terms, the economic value of the consequences arising from a roadside hazard accident are significantly less than the economic value of the consequences arising from a major road/rail accident. However, these are more than balanced by the probability of a roadside hazard accident occurring being that much greater than a major road/rail accident. The comparison of the economic risk values is shown in Figure 3.2.

Figure 3.2: Comparison of the Economic Risk Values for a Typical Road Hazard, Typical Off Road or Third Party Hazards and Railway Lines (see Annex 5 Section 5).

3.14 The comparison suggests that unprotected roadside hazards alongside English major roads have a greater economic risk value than road/rail hazards where a road vehicle may reach a railway line, despite the possible disproportionately high consequences of such an event.

3.15 The Group also discussed the implications of the rail industry's approach towards the causes of major accidents (that is, accidents involving multiple passenger fatalities). The rail industry takes all reasonable practicable measures towards unsafe acts and conditions, which are in the direct control of the rail industry, to reduce the chances of such accidents occurring. They have a long term goal to reduce injuries from train accidents near to zero by 2009.

3.16 All forms of transport involve a degree of risk. As with other modes of transport and other industries, the most effective ways of increasing railway safety involve driving down the number of events and procedures that can represent a serious risk.

3.17 Highway authorities, also take all reasonable measures to reduce accidents, but are in a different position because they exercise less control over the actions of drivers on their network. It is therefore recognised that the scope to reduce risk is different between road and rail. The Group recognised this position but considered that the rail industry's approach should be taken into account when deciding on the level of highway protection at road over rail locations.