Radstats: Child Pedestrian Accidents in the UK, Monitoring the UK Government's Road Safety Strategy

Radical
Statistics

Child Pedestrian Accidents in the UK, Monitoring the UK Government's Road Safety Strategy

Paul Hewson

Introduction

In terms of crude league tables of casualties, the official view is that the UK compares well in relation to most industrialised countries in terms of the overall numbers of casualties injured in road accidents. In contrast, there is evidence that Britain has a less satisfactory record when child pedestrian casualties are considered.

Table 1. Child (0-14) pedestrian fatality rate per 100,000: 1997

Country

Child Pedestrian
Fatalities per 100,000

Great Britain 1.21

Austria 0.79

Belgium 0.94

Denmark 0.85

Finland 0.94

France 0.91

Germany 0.64

Ireland 1.31

Italy 0.49

Netherlands 0.66

Norway 0.81

Spain 0.94

Sweden 0.54

Switzerland 0.96

Source: RSAP(00)2 DETR Road Safety Advisory Panel Papers: Improving Child Road Safety - Options for Next Steps http://www.roads.dtlr.gov.uk/roadsafety/strategy/rsap/03/htm

This picture has been well established for some time (Plowden and Hillman, 1984; Davis, 1992) yet there are several implications of official complacency. The DETR's 1998 "factsheet" on child road casualties makes no mention of the adverse comparison with the remainder of Europe (DETR, 1999), but rather concentrates on the reductions in the number of recorded casualties over time. The paper submitted to the Government's Road Safety Advisory Panel (RCAP, 2000) describing the development of a time series model forecasting the strategic casualty reduction targets for 2010 states:

The original PC spreadsheet did not forecast child casualties per se since the concept of a separate casualty reduction target for children arose at a late stage in the political process of developing the new road safety strategy.

A year later the DETR (2001a) stated, in a press release for a literature review on the involvement of children belonging to ethnic minority groups in road accidents:

The UK has the best road safety record in Europe.

It may well be that the UK does indeed have the best road safety record in Europe if road safety is taken to mean numbers of recorded casualties and if Sweden is excluded from the definition of Europe. This statement however, on an overall road safety record, seems somewhat misplaced in a press release relating to child pedestrian accidents where the analogous record for Great Britain appears to be amongst the worst in Europe.

This article aims to review the UK government's current road safety strategy as it relates to child pedestrians, and to consider how these strategies relate to recent literature. Road safety is a well-studied area of transportation policy, and there was considerable optimism that strategies might now be based on research evidence rather than free-market ideology. As with much current governance in the UK, delivery is directed towards easily measured targets. It will be argued that road safety is one example of an area of service delivery where target chasing does not adequately address the issues and may in fact be counter-productive. For example, Richter et al (2001) have argued in the U.S. context that lower casualty figures actually represent a failure of policy, and were held to be more attributable to the unacceptable lifestyle changes forced on many sections of the community than to delivery of a road safety policy. Perhaps partly due to the complacency associated with the optimistic view of casualty counts, one other aspect of the Road Safety Strategy is the boldness with which it is enacted. Plowden and Hillman (1984) argued about the dangers of timidity, yet timidity in the face of the car lobby seems to be a dominating feature of the current approach to transportation policy (Begg, 2001).

If delivery of Government Policy has to be managed by performance targets, it will therefore be argued that it is necessary to develop measures of "road danger" and to monitor reductions in this concept as well as casualty reduction. Even if freed from the requirements of performance management, it is vital that there is more widespread official recognition of the difference between "road danger" and counts of recorded casualties.

Road Safety Strategy

The Government has set out a broad strategy for the next decade (DETR, 2000b), which includes casualty reduction targets. These headline targets include the aspiration of a 50% reduction in the number of children killed and seriously injured by all modes of transport. In so far as it relates to children, the bulk of the strategy concentrates on education and therefore appears to be aimed very much at reducing the number of child pedestrian casualties. The strategy however includes some mention of 20mph zones, Home Zones, increased speed enforcement and other engineering approaches. However, there remains a strong emphasis on education. This is interesting given that the government's own briefing paper (RCAP, 2000) refers to different land use patterns and the existence of less traffic calming and other measures to slow traffic as possible reasons for the relatively poor child pedestrian record in the UK. Unfortunately, in many spheres, "road safety" has become synonymous with education of children, and the perception of success or failure of past efforts depends entirely on presentation of statistical information. Superficial presentation of the numbers of child pedestrian casualties reported to the Police, which have fallen consistently for decades, is taken to support the view that education efforts in the past have been successful. The converse view could consider that road safety education conspires with increasing traffic levels in forcing unacceptable lifestyle patterns on children. It can therefore be suggested that complacency in marshalling headline performance indicators relating to road casualties is counter-productive for the very group they are intended to serve. It is particularly counter-productive in the mainly low-income households who do not have access to a car. If it is agreed that modal shift into cars is the predominant factor contributing to lowering road casualties in the UK to date, those households lacking income for this action can reasonably be expected to suffer higher casualty rates. It is also particularly pernicious in relation to child pedestrians whose requirements go far beyond the need to be conveyed between locations (Millward and Wheway, 1997). This modal shift is invidious in creating an attitude that cars belong on the road, pedestrians (including children playing) are then expected to be elsewhere. One example of the way in which this phenomena is realised is given by Malek et al (1990). They initially cite well established work suggesting that in 90% of child pedestrian accidents the children have attempted some kind of evasive manoeuvre, yet in only 10% of accidents has the driver attempted any evasive manoeuvre. They go on to report evidence suggesting that when drivers attempt some kind of defensive manoeuvre, there is strong evidence that the accident severity is reduced even if not avoided.

Apart from the specifics of the road safety strategy, it had been hoped that other areas of transport policy, such as the commitment to integrated transport and the Road Traffic Reduction Act of 1997 would have a positive impact, however powers in this Act have yet to be exercised. Begg (2001) suggests that transport policy is lead by groups campaigning for the perceived needs of the motorists.

Education

It has been argued that there is excessive emphasis on child education in the road safety strategy. Recent research that has a bearing on this view includes the RoSPA claim that children under 8 lack the necessary skills to interact with traffic safely (Rogers, 1993), supported by others such as Schieber and Thompson (1996). Connelly et al (1998) noted that children cannot accurately judge gaps in the traffic (pertinent in a heavy traffic environment common in the UK). The firmest limit placed on the role of education is given by McGregor et al (1999) who suggest that children's accidents arise not because of a lack of knowledge, but because children make "mistakes". As they argue, the answer to this cannot be better education, it can only be to work towards a more forgiving traffic environment. Despite this evidence, large components of the Road Safety Strategy put the emphasis for safety of this group on successful uptake of training. The strategy (DETR, 2000b - below) sets out a large number of educational activities aimed at child pedestrians.

It must be noted that School Travel Plans offer far more than just education. It is envisaged that there will be commitment to infrastructure improvements by the local authority, and they do offer some limited scope for approaching road safety in a partnership manner. A community based scheme could require drivers (at least those responsible for the children) to consider their responsibilities in terms of the road safety of others. It could also raise awareness of the rights of pedestrians and could be valuable as a long term investment promoting culture shift. However, the vast majority of the Road Safety Strategy represents an incremental approach to the education of children to make them "safer" road users, an approach that has been in use for decades.

UK Road Safety Strategy: Components relating to Child Education

Issue (with RoSPA) national guidelines on roadside pedestrian training

Continue to encourage schools to teach road safety in PHSE at key stages 1, 2 and 3
Work with road safety professionals to ensure road safety resources are used in schools
From September 2000 a new road safety aid will be made available for classrooms
Produce road safety education guidelines for schools and professionals
Encourage parents to get involved through ``one step ahead'' for new parents
Encourage parents to get involved through resources for the transition from primary to secondary school
Explore potential for practical road safety training in play schemes and homework clubs
Increase access to road safety materials through the internet
Make "Rosalind" more user-friendly (database of road safety education linked materials)
Discuss with road safety professionals and community groups how to encourage people to start up and run training schemes
Develop a programme promoting child pedestrian training in deprived areas
Monitor high risk groups and explore ways of improving their safety.
Support school travel plans
Raise the profile of school travel issues and spread best practice through the School Travel Advisory Group.

Relationships between accident rates and exposure.

Despite the apparently positive indication of a fall in child pedestrian casualties, it is difficult to be unaware that child freedom in has been increasingly restricted in the last few decades (Hillman et al 1990). Roberts (1995) states the claim very clearly that child accident rates have fallen because of children losing independent mobility. Whilst it is possible to estimate the number of vehicle kilometres travelled and calculate data for car accident rates for different road types and locations, it is difficult to arrive at an analogous measure of child pedestrian exposure. One recent attempt to enumerate the effect of modal shift (DiGuiseppi et al, 1997), by analysing available official statistics, suggested that the apparent decreases in child pedestrian accidents were at least in part attributable to a decrease in overall distance walked. Ironically, as part of this decrease has been facilitated by a modal shift to motorised transport, it was also suggested that the decrease in motor accident death rates was to some extent nullified by the increase in the average number of miles travelled by this mode. This implies that modal shift is not risk free in terms of accidents. DiGuiseppi et al, (1997) further reported that in 1985 average distance walked by a child aged 0 - 14 was 247 miles, in 1992 this had fallen to 197 miles (a decline of 20%) whereas average distance travelled by car increased from 2259 miles to 3158 miles (40% increase). This paper reported that there had been a more pronounced modal shift in girls than in boys which may indicate some parental risk confounding with "Stranger Danger". In addition to a suggested reduction in child deaths population of 37% (with confidence intervals estimated between 27% and 45%) the paper suggested that there had been a reduction in child accident rates per mile walked of only 24% (with 95% confidence limits estimated between 12% and 34%).

It is useful however to consider carefully both the numerator and denominator in these rates. Responsibility for official recording of road traffic accidents lies with the police, there is plenty of work to suggest considerable under-reporting of pedestrian accidents (Keigan et al, 1999). Whilst estimates of injury accidents may vary by a factor of seven, it is hoped that the police can cope with the task of recording child pedestrian fatalities accurately. Estimates of walking are based on the National Travel Survey and are obviously problematical in terms of estimating child exposure to traffic, not all of which comes from a recordable journey. Child exposure may have been reduced by considerably more than the estimates suggested by the National Travel Survey, in which case the accident "rate" figures are more worrying than those presented, or child exposure may not have altered at all. There also seems little reason to believe that a percentage reduction in recorded miles walked has any bearing on a percentage reduction in accidents and that the rates are comparable.

Arguments have been made for survey information, analogous to the British Crime Survey, to provide a validation of the numbers of officially recorded accidents (Plowden and Hillman, 1984), however some crude confirmation of the scale of under-recording is available when considering Department of Health Data for In-patient episodes. The most recent published figures are for 1999 / 2000 (DoH, 2000). It needs to be stressed that these data should only record a small subset of those accidents recorded by the police, and that there are a number of problematic mismatches in this comparison. Firstly, not all "serious" road accidents will lead to an in-patient episode. The STATs 20 document (DETR, 2000a) which sets out the data recording requirements for serious casualties is unclear, but lists injuries such as "fracture", "severe cuts" and others which may not necessarily lead to an inpatient episode. The DoH data presented is for England alone, and reports for the 0 - 14 age group and is presented for the financial year 1999 / 2000. The DETR data however is for Great Britain, broken down for 0 - 15 year olds and for calendar year 1999. However, according the DETR 107 child pedestrians were killed, 3350 seriously injured. According to the DoH provisional figures for financial year 1999 / 2000 for England alone, there were 3700 in-patient episodes involving child pedestrians aged 0 - 14 from pedestrian involvement in road traffic accidents. Whilst direct comparison is not possible, and it would not be sensible to draw too many conclusions without further effort to match the data, it is clear that there is considerable under-reporting of serious accidents to child pedestrians, a category generally held to be fairly reliable. It is noteworthy that the STATs 19 data is recorded by the same agency against which performance targets have been set for reducing the number of serious accidents. Plowden and Hillman, (1984) were calling for the medical profession to assess the severity of casualties almost two decades ago. This could provide a check both on the police under-rating the severity of the casualty and secondly in ensuring that more accidents which currently go unreported to the police can be officially noted. The police reporting system is currently subject to a regular quality review, it will be interesting to see if these recommendations will be followed now that the technology is readily available.

It is possible to evidence worrying official support for the use of crude casualty counts from the STATs 19 data where there has been insufficient consideration of context or exposure. For example, all authorities have five years to fundamentally review all their activities, and as with education and social services these are then liable to inspection. The "Best Value" Inspection of Lincolnshire (BVI, 2001) commended the authority for its low numbers of pedestrian accidents. This commendation was made in the absence of any data on the amount of pedestrian activity, indeed, given the high rate of car ownership in the county alluded to in the same report it is quite plausible that pedestrian activity is comparatively low in the county. The Best Value Inspectorate's belief in the Performance Indicators is even more misplaced than that used for mere comparison of performance; paragraph 40 recommended that the road accident database be used to plan routes for activities such as school buses and meals on wheels routes. Fortunately no recommendations were made in respect of planning activities that would impact on child pedestrian activities, such as trips or walking buses. Nevertheless this seems like a worrying misunderstanding of the accident data on the part of the body charged with judging the effectiveness of local authorities.

'Model' for child accident rates

DETR have a 'model' for child accident rates (Johnson et al, 2001):

Time on Street x Degree of hazard presented by the environment

Ability to deal with risk

One of the refinements to previous educational work is that the government intend identifying groups of children most at risk, and ensuring that they are trained to use the road safely. However, it is striking in the above "model" that the degree of hazard presented by the environment is an aggregate level concept, whereas the other two are individual level effects, and even more striking that given suitable data, statistical models to resolve these are commonplace. Many of the risk factors described may affect either level of this model. Additionally, epidemiological work described below clearly suggests that "Time on Street" is a massive simplification of a more complex exposure concept.

Having identified groups of children at risk (rather than the areas in which they live) does not explicitly lead to a solution. Even accepting considerable variations, even systematic variations, in the denominator "ability to deal with risk" does not evade the possibility that accident rates can be reduced now. With little further research it is clear that strategies exist to reduce the degree of hazard presented by the environment. It could be argued that given a diverse population, we should accept a wide range of "ability to deal with risk" and develop complementary solutions that would benefit all children, at the same time as trying to enhance children's ability to deal with risk. In addition, it is a concern that it is easier to frighten parents into reducing "Time on street" rather than to increase a child's "ability to deal with risk" or to decrease the "degree of hazard presented by the environment". All will reduce child casualties, adopting a lifestyle which minimised exposure is however is detrimental to both the child's social development and also the communities in which they live.

Individual ability to deal with risk

Johnson et al, (2001) reviewed a wide range of published studies examining individual proneness to road accident involvement over the last forty years. Potentially relevant social and psycho-social factors reviewed in this report that have been raised at some point include family size, birth order, female head of household, maternal education level, inability to speak English, indices of child stress or family stress, divorce or marital discord, physical or mental illness, poor or ineffectual child-rearing, atypical lifestyle, maternal pre-occupation (either through work or pregnancy). The list of factors reveals as much about government research prejudices over the last four decades as it does about factors which may have some bearing on child pedestrian accident rates, for example do the studies examining the mothers' role reflect on lower car ownership and driving by females or a more significant role in child-rearing. More recent work (Christoffel et al, 1996) suggested a number of specific factors, children with higher aggression, hyperactivity and anti-social behaviour ratings experienced greater risk, as did children from families with higher family stress and lower levels of family supportiveness and cohesion. This identifies children who may exhibit more risk-taking behaviour, but whether this associates with low socio-economic status or ethnic grouping is questionable. Christie (1995) also identifies a number of individual risk factors associated with accidents, however many of these can be taken as consistent with belonging to a family where a car is available and used to meet all the transport requirements a child is allowed. White et al (2000) have suggested that areas with high child pedestrian accident involvement tend to have low car ownership rates. Both pieces of work tend to support the view that modal shift into cars is responsible for the declining numbers in child pedestrian casualties, rather than effectiveness of any education programmes.

DETR have recently made money available for specifically targeted road safety training (DETR, 2001c) but this raises two questions, firstly, as to whether the targeting is accurate and secondly as to whether this is an exercise in "victim-blaming". In targeting deprived areas it certainly creates the illusion of equity but whether children in low socio-economic areas need training or whether the drivers who drive through low socio-economic areas need training is an open question. One consequence of urban planning is that vast numbers of drivers travel through many deprived areas because this is where major arterial routes around cities are located. It therefore follows that a case can be made for accepting diversity in children's ability to deal with traffic, and unifying urban planning and remedial engineering on the assumption that road safety is a condition where danger is reduced, rather than judging success solely on casualty counts.

Immigration and accidents

One of the major research commitments of the current government has been to consider the involvement of children belonging to ethnic minorities in pedestrian accidents. This differential was observed over two decades ago but it is still unclear as to whether there are ethic group differences, socio-economic group differences or area of residence differences. Work containing some incredibly sweeping generalisations (Jacobs et al, 1981) suggested that drivers in developing countries were far less disciplined than in the U.K. particularly in relation to pedestrian facilities. Despite this the DETR (Johnson et al, 2001) suggest that the reason ethnic minorities suffer higher accident rates may be due to unfamiliarity with traffic conditions in the UK and failure of parents to adapt. This view of immigration as a risk factor is not universally supported (Abdalla, 1997). One recent counter to the view of immigration as a risk factor can be easily considered in a UK context, in that tourists have to deal with the UK's relative novelty in driving on the left side of the road. This should provide evidence for the significance of an "unfamiliarity" domain in differential road accident involvement rates. Scottish Office evidence for car drivers suggested that tourists (mainly American and German and therefore used to driving on the opposite side of the road) were not over-represented in the accident data when their mileage was considered (Sharples and Fletcher, 2001). Whilst it is not clear that the same can be claimed for child pedestrians, no evidence is available to suggest that unfamiliarity with a different road system is any worse than difficulties coping with a more familiar set up.

It has been suggested (Christie, 1995) that "non-white" parents have lower risk perceptions than the "white" groupings. Given the incredible diversity in this "non-white" group it is questionable whether an overall average risk perception has much validity. Indeed, given the ability of the population to handle the concept of risk it is debatable whether attempts to enumerate parental assessments of risk are likely to be reliable in any group. Samuels (1993) describes "Simpson's Paradox", a well recognised statistical phenomena whereby simplifying conclusions drawn from aggregates of diverse sub-populations can be seriously flawed. However, even in the absence of accurate assessment of risk, if the only solution is removal of children from risk this may be an unjust solution to the problem. The over-involvement of children of families who have recently immigrated is only relevant if one holds the child responsible for the accident. If one takes the view that government policy should make the environment safer for all children, identifying those groups most at risk is less important than developing a safer environment, unless it helps prioritise those areas which need more urgent remedial action.

Engineering

Whilst it will be argued that child pedestrian road safety has to be dealt with in a planning and land use context, all the current substandard roads will need to be developed in such a way as to allow passable levels of child safety. A number of engineering components exist within the Road Safety Strategy.

UK Road Safety Strategy: Components relating to Engineering (where most directly relevant to Children)

Local Authorities must consider using their powers to create more 20mph zones (with suitable traffic calming and parking restrictions around schools and in residential areas where most child accidents occur)
Local Authorities must consider providing safe crossing facilities (on busy through roads where traffic calming is not appropriate)
Highways Agency to develop and implement child friendly areas on trunk roads near schools and in residential areas
Supporting local authorities putting in place Home Zone schemes (to slow vehicles down and give more priority for walking and cycling. (The latter comprises a pilot of eight zones with a further nine selected for trials in 2000). Support school travel plans

Evidence for the need of 20mph limits (or lower) in residential areas seems incontrovertible. An editorial in the BMJ (Pilkington, 2000) argued strongly for the introduction of more 20 mph limits and the resulting debate suggested a good degree of support. Work by the Scottish Office evaluated their very limited introduction (Burns et al, 1999). These claimed successful introduction of such schemes both in terms of speed reduction, accident rates and increased resident perception of safety. 20 mph zones are a cheap way of highlighting to consenting drivers the need for more careful speed, unfortunately traffic calming (Ross Silcock, 1999) and enforcement may still be necessary, especially on through routes. Kraus et al ((1996) have enumerated the risk to child pedestrians of speed, the general arguments for a review of speed policy generally are set out by Plowden and Hillman (1996), little more will be said except to briefly review enforcement approaches later.

It is interesting that while inner city communities are frequently divided by road development (for example the current extension of the M74 in Glasgow) "stress" caused to suburban communities caused by car dependency has also been noted (Gwilliam et al, 1999). It seems likely that this stress will continue to grow, possibly to the point that many of the environmental advantages in respect of child road safety of more affluent areas (more cul-de-sacs, less kerb parking, less traffic) will be negated. Epidemiological work can confirm the risks posed by the environment for children. A study reported in the BMJ described careful work examining child exposure and accident rates. A case-control study carried out in New Zealand suggested that children living in high traffic neighbourhoods were exposed to 13 times the risk of children in low traffic neighbourhoods (Roberts et al, 1995). Kerb parking was also a significant factor, as was speed (although this work identified the middle speed sites as the riskiest, perhaps because attempting crossings at the highest risk sites was impractical). The paper pointed out that these factors have to be addressed at a government level rather than at a parent / child level.

Christie (1995) noted that pre-1914 estates and through roads were significant risk factors, although she contradicted earlier work in not finding street parking a risk factor. Parking however was not specifically controlled, so the higher risk of fast through roads with no parking may have overwhelmed the evidence of kerb parking on residential streets). Stevenson et al (1996) reported on case control studies in Australia and found speed (when controlling for volume of traffic) and kerb parking to be significant risk factors. Bly et al (1999) compared child exposure in three countries, Britain, France and the Netherlands. Their work suggested that there was little difference in the amount of time children spend playing near roads in the three countries and that children in Britain crossed the road less often. The report also suggested that due to land usage patterns in Britain, children crossed more major, heavily trafficked and wider roads. Also, British children were more likely to be accompanied by other children (rather than by adults), to use unmarked crossings or to cross mid-block. This all points strongly to the need to consider land use and planning in order to make a significant difference to child pedestrian safety. This has been the case in many countries in Europe, such as Sweden, and has been particularly successful. Towner et al (1996) reported an attempt to enumerate children's exposure to various kinds of hazards on their journey to secondary schools and this could form the basis of methodology to identify priority areas for action.

Community involvement.

Stevenson (1999) report results from community interventions, where in addition to environmental work, a full programme of community involvement was tested and found to considerably enhance the success of any interventions. It has been mentioned above that the School Travel Plan offers some potential in this regard, but children's needs go well beyond the need to travel to one fixed destination. The need for play is well recognised. Joseph Rowntree Foundation research (Millward and Wheway, 1997) suggested that traffic speed on estates needed to be lowered to 10mph. In terms of planning, the report suggested that children enjoyed access to a large outdoor environment ideally served by good footpath network and cul-de-sac layout. Many of our more deprived areas date from a period when this type of urban design was not considered, and efforts therefore need to be made to "engineer" an environment as close as possible to the more desirable environments. The report also emphasised the extent to which children enjoy their mobility for different types of play and social contact. The strong suggestion was that restricting children to "safe" designated play-areas was not desirable and ultimately impractical, which suggests that any amount of training to find "safe areas to play" will have limited impact. In Germany however, von Kries et al (1998) did find evidence that safe areas to play were beneficial, but also highlighted the need for crossing facilities and speed control. This may suggest that attempts by the Highways Agency to find "safe areas" could have limited effect if implemented in isolation.

Beyond children's needs to play, there are general benefits to be derived from more appropriate use of urban space. For example a Joseph Rowntree Study on deprived parts of Nottingham (Silburn et al, 1999) emphasised the importance of transport links and proximity to the town centre as significant positive components enhancing quality of life in the neighbourhoods studied. Links which allow some child freedom within this, rather than dependency on a car and therefore on adults have to be valuable rights that need defending. Piloting 14 Home Zones within the UK is clearly a start, but given the plentiful data from the continent on all the benefits and an understanding of potential pitfalls this seems rather modest. Home Zones currently have an uncertain legal status in the UK, a situation that could be very simply resolved. There also seems little reason for not adopting best practice in terms of new developments. Preston (1995) argues strongly, using the available evidence base, for the need for Home Zones. This remains a timid area of implementation, subject to tight central control and successful completion of business cases to obtain the necessary finances. Community involvement seems essential in the UK given the prevailing car-culture. Roberts and Coggan (1994) review the way the onus for accidents is shifted onto children. A Home Zone program, with the law changes used in parts of continental Europe represent the first step in encouraging responsible car driving, and removing the onus from children who particularly when young, are considered to be inherently incapable of interacting with traffic safely.

Enforcement

There is one item in the Road Safety Strategy dealing with enforcement.

UK Road Safety Strategy: Components relating to Enforcement (where most directly relevant to Children)

Local authorities must consider making enforcement more effective (especially by the use of cameras on busy roads where many accidents happen)

Currently, the Government is implementing a revenue hypothecation scheme, whereby partnerships of local authorities and police forces can apply to keep the revenue from speed fines to purchase more speed enforcement. The role of speed in accidents is well documented, and Plowden and Hillman (1996) have published a clear and costed account of the role of speed reduction in reducing accidents which would be beneficial for all road users. Despite the strong evidence of the role of speed, the only enforcement approach is in terms of self-financing speed cameras. Some research continues into in-car speed regulators, European Directives on Commercial Vehicle limiters are currently being considered and some effort is directed towards limiting advertising that makes speed look attractive. As Plowden and Hillman (1996) argue, speed cameras are a small component in the approach to enforcing speed limits. Unfortunately, there is little cause for optimism that the government will follow the strong evidence base on the need to rigorously enforce speed limits (Begg, 2001). It is noteworthy that the EU, which has the financial muscle to impose its will, has decided that directives for more pedestrian friendly cars will be of a voluntary nature only.

It may be instructive to review some official statistics describing police activity in relation to traffic are published by the Home Office (Wilkins et al, 2000). The influence of the introduction of speed cameras has had a clear effect on police activity against this offence. Whilst speed of traffic is a risk factor (Roberts et al, 1995), the bigger question remains whether speed limits are appropriate where children require mobility, for example it is possible to question how often speed cameras are located in residential areas.

Click HERE for Figure 1.

The series relating to "Obstruction, waiting and parking offences" has been plotted on the right hand axis, the increase seen in 1998 and 1999 is due to transfer of powers to a few local authorities who processed nearly 4,000,000 offences each year. The increase in "Miscellaneous'" represents seat belt offences.

Remarkably, policing of parking offences declined dramatically until these powers were transferred to local authorities. Given that even legally parked cars are a risk factor, it seems unlikely that the effect of illegal parking on pedestrian safety is negligible. It is also apparent that the number of offences relating to signs and pedestrian rights has been declining, yet there is evidence that up to a third of child pedestrian accidents happen on pedestrian crossings or on pavements (Davis, 1992).

Monitoring policy delivery

Research commissioned by the DETR (Bly et al, 1999) specifically highlighted the use of unmarked or mid-block crossings, in addition to land and road usage patterns which may contribute to the differential child accident rates in comparison with other countries. This seems like a simple and uncontroversial area which could be remedied quickly. It is therefore instructive to see how Audit Commission Performance Indicators have been used as a means of comparing the performance of different authorities. The Indicators for 1999 - 2000 give a good idea of official priorities.

Audit Commission Performance Indicators 2000 - 2001 (some of these indicators continue after 2001 under a different name).

P1: The % of streetlights not working as planned
P2a: The % of repairs to dangerous damage to roads and pavements which were carried out within 24 hours
P2b: The number of serious accidents per 1,000,000 miles travelled by a vehicle on principal roads
P2c: The % of principal roads which have reached the point at which repairs to prolong their future life should be considered
P3: Number of days major council roadworks were in place per mile of busy road
P4: The cost of highway maintenance per 100 miles travelled by a vehicle
P5: The % of pedestrian crossings with facilities for disabled people
P6a: The % of footpaths and other rights of way which were: signposted where they leave a road
P6b: Easy to use by members of the public
P7: Net spending per head of population on public transport

Best Value Performance Indicators.

BVPI 1999 Number of road accident casualties per 100,000 pop broken down by (i) nature of casualties (ii) road user type. Casualty categories: (a) killed/seriously injured (b) slight injuries. Road user types: (a) pedestrians (b) pedal cyclists (c) two wheeled motor vehicle users (d) car users (e) other vehicle users some authorities also break this figure down by child / adult.<

Local Performance Indicators Proposed 2001 onwards (voluntary)

The % of road traffic accidents involving death or personal injury in which at least one driver tested positive for alcohol
Percentage of primary schools with school crossing facilities
Percentage of primary school crossings fully staffed during school terms
Percentage of primary school crossings at or above satisfactory level in Health and Safety inspection
Percentage of primary school pupils receiving road safety advice during the year
Percentage of children between x and y years of age achieving their cycling proficiency test in the year
Percentage of signal controlled junctions with green man facilities

There is no national requirement to monitor crossings other than recording the percentage which are easy to use by disabled people, an indicator that can be "improved" by removing crossings as much as by upgrading them. Emphasis on School Crossing Patrols is also interesting as this seems to reinforce the idea that the roads are "too dangerous" for children to be allowed near them unsupervised, and again, only provides protection on the journey to school. It is also clear in the national indicators that avoiding disruption warrants a league table type performance indicator, whereas monitoring how often traffic signals have a pedestrian phase is a voluntary indicator considered by only a few local authorities. The differential level of skill required to cross a road with and without a "green man" can be challenging for fit adults. Finally, the indicators give an input measure of child road safety education. If the government is serious in its belief that children from lower socio-economic status groups and ethnic minority populations are at enhanced accident risk then this indicator should at least be weighted to allow an emphasis on educating those people most in need. Conversely, there is little to reflect the engineering or adoption of a planning strategy that most available evidence suggests is most needed. It is also noteworthy that the DTLR publish speed surveys regularly, but this information does not appear to be monitored as an intermediate stage in achieving "road safety". This perhaps confirms Plowden and Hillman's (1996) view that there is insufficient commitment to tackling the problems of speed.

Current Official Research Priorities

Reference has been made to the wealth of available literature in terms of road safety generally, but delivery in some areas is still slow due to the need to "evaluate pilot schemes". It is interesting to examine current research priorities, and to compare these with areas where delivery is restricted. £4.500m excluding VAT was allocated by DETR (DETR, 2001d) for road safety research. The designated priorities were:-

Research work funded by DETR in 2000 - 2001 included:

Safety of vulnerable road users (Cycle training, ethnic minority children, child pedestrian sure), application of safety measures;
Driver behaviour (learner and novice drivers, evaluation of driver improvement schemes, motorcycle safety);
Impairment (the effects of drugs on driving, police enforcement strategies, field impairment test training for police officers, driver fatigue counter measures, sleepiness and moderate alcohol intake, effects of over the counter medicines on driving, the ageing driver);
Speed management (signing on rural roads),
Review of values of prevention of road accident casualties;
Fitness to drive, (cardiovascular risks, epilepsy and cerebrovascular disease, diabetes and psychiatric conditions).

No finance is committed towards researching Home Zones, or to enforcement of speed limits, yet implementation of initiatives in this respect remains relatively timid. It is notable that research continues into application of safety measures when non-quantitative evidence already suggests that those available are not used correctly (BBC 2001) and when it can be strongly argued that these falsely shift the onus from the motorist (Davis, 1992) to the vulnerable road user. It is also significant that a purely notional exercise, the review of the values of prevention of road accident casualties, retains any serious investment. Full implementation of a Home Zone programme within Great Britain may require little further research given the wealth of evidence available from the continent, but clearly the UK context for legislation needs work. Home Zones currently have no legal definition. One rather high profile case involving a child cyclist (CTC, 2001), where the child's parents were nearly sued for letting their child on the road, clearly demonstrates that in a litigative culture children's' rights to be on the roads need to be enshrined in law.

Child Safety Audits and the need for cross-profession dialogue

Finally, there is one broad avenue which allows creative scope for local authorities to make serious inroads into child pedestrian safety, the requirement to conduct a child safety audit. No guidance has yet been issued centrally, so it is possible that there is still opportunity to influence the scope of this audit and move away from a reported-casualty led definition of road safety. An over-riding concern in relation to road safety is that it may have become an over-specialised field, reluctant to learn from outside its own spere (Davis, 1992). A B.M.J. editorial (not a journal noted for environmental or political fundamentalism) in 1995 (Roberts, 1995) posited that "The key to prevention is a change in transport policy". In reporting on his paper in that issue it lamented that despite 20 years detailed study of child accident mortality, epidemiological methods had rarely been applied to road casualty data. The editorial commented on the link between increasing traffic and decreasing pedestrian mobility and suggested that professionals concerned with child mortality had a role to play in a transport agenda dominated by the motor lobby. There was a very strong suggestion in the editorial that the medical and road safety literatures had become separate specialised fields, and opportunity for an informed dialogue that could have enriched both fields had been missed. These comments are of huge importance given the forthcoming "Child Safety Audits" which Local Authorities are required to conduct. There are many epidemiological and sociological methods available, and this represents a real opportunity to resolve questions surrounding whether risk factors are child based or area based. The answer to these questions can provide strong evidence as to whether the emphasis in child safety should be on engineering and planning rather than child education.

Conclusions

Motorways can be deemed "safe" in that child pedestrian accidents are reported only rarely, but in common with urban settings the apparently low accident rate could well be due to low exposure to risk. And if low exposure in an urban context is at the expense of play, walking and cycling with all the concomitant health penalties and social development penalties claiming success for the UK road safety record is at best disingenuous. Accepting the views of Richter et al (2001) supports the opinion that accident rates are falling due to modal shift rather than government intervention. Clearly, those who cannot afford modal shift into the car are going to be disproportionately affected. If one takes the view that there is work that government (including central, local and law enforcement components) could be doing to address the situation this represents a significant cause of institutional discrimination with acute consequences for those directly affected. Whilst the role of modal shift in reducing casualty rates is universally agreed, neither is it yet clear whether apparent differences in casualty rates are due to individual factors (lifestyle or ability to deal with traffic) or the area in which you live. There is overwhelming evidence pointing to the importance of urban design, an issue that only government can address. Regardless of whether there are differences between different groups in terms of their ability to deal with traffic, which would validate the government's attempts to level the playing field with child pedestrian training, it can still be argued that a goal of transport policy should be to make the environment safer for children of a wide range of abilities. This transport and urban design policy should accept the diversity in children's road sense and accommodate it. Given the cost of owning and running a car are not trivial it should be accepted as a condition of a driving licence that there will be vulnerable road users, that children will play in residential roads, and that drivers are responsible for taking all reasonable care to avoid injury in these situations.

The overwhelming obstacle in any debate on child pedestrian road safety is that road danger cannot be easily quantified. Whilst some attempts have been made to review child exposure there is still excessive official reliance on the numbers of casualties reported to the police and, as with many other areas of governance, targets have been set for casualty reduction. However, despite the targets, it can be argued from the available evidence that the government has set out relatively little that will aid in achieving this target. On the basis of past trends, the target is easily achievable assuming ongoing modal shift. Unfortunately, the modal shift into cars may be nearing capacity, and so much more ambitious policies may have to be implemented to meet this target. In any event, in an environment that dictates that you can only govern what you can measure, different performance indicators will be needed to monitor road safety. In addition to the casualty reduction targets, the government should set "road danger" reduction targets.

Possible alternative measures for a concept of "Road Safety"

Analagous to the BCS it may be possible to collect data from survey on road accidents, near accidents and fear of accidents. This has been suggested several times, e.g. Plowden and Hillman (1984);
the age at which children are accorded free access to roads to play and travel (Hillman et al 1990)
Number of children playing in street / freely mobile along street as a proportion of child population in the area.

Alternatively, if engineering objectivity is sought it would be possible to sample

Type of road, residential or retail, through road, major urban through road, other, the number of such roads crossed on typical journeys;
Observed speed of traffic (the actual speed limit set and compliance with the limit are of little relevance from the child's point of view);
Volume of traffic;
Amount of parked cars;
Provision of facilities to cross road where necessary (in relation to the routes child pedestrians wish to take rather than a crude count of the numbers available).

It has been observed that boldness is needed in delivery of a Road Safety Strategy (Plowden and Hillman, 1984), and many authorities point to the need to consider land use and planning issues. Planning and community concepts are under-represented in the government's current strategy. The Home Zone program is long overdue in Britain, and it is regrettable that such a cautious start is being made, nevertheless it is progress in the right direction. It seems simple and cheap to convert all residential streets to 20mph or lower, and it also seems relatively simple to give Home Zones a more definite legal status. However, many entrenched attitudes to Road Safety need to be shifted. The Child Safety Audit, if developed correctly, offers much scope for Local Authorities to influence the definition of the problems surrounding child pedestrian safety. The environmental imperative surrounding this issue is obvious to many, but the issues relating to social development, equity and health are now very well rehearsed.

Acknowledgements

I would like to that the referees for their forbearance in reviewing earlier versions of this piece of work.

References

Abdalla, I.M. (1997), "Statistical investigation and modelling of the relationships between road accidents and social characteristics", PhD Thesis, Napier University

Begg, D. (2001), "Analysis: Driven to Death" (Comment and Analysis), The Guardian 30 August 2001, Page 17

Bly, P., M. Dix and C. Stephenson (1999), "Comparative Study of European Child Pedestrian Exposure and Accidents", DETR, London

BVI (2001) "Lincolnshire County Council Road Safety", Best Value Inspection Service, London UK.

Burns, A., N. Johnstone and N.Macdonald (2001), "20 mph Speed Reduction Initiative", Scottish Executive Central Research Unit: Development Department Research Programme Research Findings 104

BBC (2000) "Rise in Child Road deaths", BBC News 29th June 2000

BBC (2001), "Coroners Warning on child seats", BBC News 3rd May 2001

Christie, N. (1995), "The high risk child pedestrian: socio-economic and environmental factors", TRL Project Report 117} TRL, Crowthorne Berks

Christoffel, K., M.Donovan, J.Schofer, K.Wills and J. Lavigne (1996), "Psychosocial factors in childhood pedestrian injury: a matched case control study", Pediatrics 97:33 - 42

Connelly, M.L., H.M. Conaglen, B.S. Parsonson and R.B. Isler (1998) "Child pedestrians' crossing gap thresholds" Accident Analysis and Prevention 30(4):443 - 453

Davis, R. (1992) "Death on the Streets; Cars and the mythology of road safety" Leading Edge Press and Publicity Ltd., Hawes, North Yorkshire

DETR (1998), "Instructions for the Completion of Road Accident Reports: With effect from 1 January 1999 (STATS 20)" DETR, London

DETR (1999), "Transport Statistics: Child Casualties in Road Accidents: Great Britain 1998" DETR, London (http://www.transtat.gov.uk/facts/accident/children/child98.htm)

DETR (2000a), "Road Accidents Great Britain 1999: The Casualty Report", The Stationary Office, London

DETR (2000b), "Tomorrow's Roads Safer For Everyone", The Stationary Office, London

DETR (2001a), "Asian kids at risk on the Roads", News Release 196 30th March 2001, DETR, London

DETR (2001c), "£10 million for child road safety lessons", News Release 260, 4 May 2001

DETR (2001d), "Compendium of research 2000 - 2001", DETR, London

DiGuiseppi, C., I.Roberts and L.Li (1997), "Influence of changing travel patterns on child death rates from injury: trend analysis", British Medical Journal 314:710 - 715

DoH (2000) "Hospital Episode Statistics" http://www.doh.gov.uk/hes/

Gwilliam, M., C. Bourne and C. Swain and A. Prat (1999), "Sustainable renewal of suburban areas", Joseph Rowntree Foundation, York

Hillman, M., J. Adams and J. Whitelegg (1990), "One False Move; A Study of Children's Independent Mobility", Policy Studies Institute, London

Kraus, J.F., E. Gerken Hooten, K.A. Brown, C. Peek-Asa, C. Heye and D.L. McArthur (1996), "Child pedestrian and bicyclist injuries; results of community surveillance and a case-control study", Injury Prevention 2: 212 - 218

von Kries, R., C. Kohne and H. von Voss (1998), "Road injuries in school age children: relation to environmental factors amenable to interventions", Injury Prevention 4:103 - 105

Keigan, M., J. Broughton and R J Tunbridge (1999), "Linkage of STATS19 and Scottish hospital in-patient data - analysis for 1980-1995" Report Number TRL420, Transport Research Laboratory, Crowthorne, Berkshire

MacGregor, C., A. Smiley and W.Dunk (1999), "Identifying Gaps in Child Pedestrian Safety: Comparing What Children Do with What Parents Teach" Transportation Research Record, 1674: 0724

Malek, M., B. Guyer and I. Lescohier (1990), "The epidemiology and prevention of child pedestrian injury", Accident Analysis and Prevention 22(4):301 - 313

Millward, A. and R. Wheway (1997), "Facilitating play on housing estates", Joseph Rowntree Foundation, York

Pilkington, P. (2000), "Reducing the speed limit to 20 mph in urban areas",

British Medical Journal 320:1160

Pitt, R., B. Guyer, C-C. Hsieh and Marvin Malek (1990), "The severity of pedestrian injuries in children: an analysis of the pedestrian injury causation study", Accident Analysis and Prevention, 22(6):549 - 559}

Plowden, S. and M. Hillman (1984), "Danger on the Road: The Needless Scourge", Policy Studies Institute, London

Plowden, S. and M. Hillman (1996), "Speed Control and Transport Policy", Policy Studies Institute, London

Preston, B. (1995), "Cost effective ways to make walking safer for children and adolescents", Injury Prevention, 1:187 = 190

Richter, E.D., P. Barach, E. Ben-Michael and T. Berman (2001), "Death and Injury from motor vehicle crashes: a public health failure, not an achievement", Injury Prevention 7:176 - 178

Roberts, I (1993), "Why have child pedestrian death rates fallen?", British Medical Journal 306:1737 - 1739

Roberts, I. and C. Coggan (1994), "Blaming children for child pedestrian injuries, Soc.Sci.Med. 38(5): 749 - 753

Roberts, I. (1995), Editorial, "Injuries to Child Pedestrians", British Medical Journal 310: 413 - 414

Roberts, I., R. Norton, R. Jackson, R. Dunn and I. Hassall (1995), "Effect of environmental factors on risk of injury of child pedestrians by motor vehicles: a case control study", British Medical Journal 310: 91 - 94

Rogers, D. (1993) "Child safety - are we doing all we can", Safety Education, Summer: 10 - 12

Ross Silcock Limited and Social Research Associates (1999), "Community Impact of Traffic Calming Schemes", Scottish Executive Central Research Unit; Development Department Research Programme, Research Findings No. 68

RSAP (2000), "Improving Child Road Safety - Improving Options for next steps", Department for Transport, Local Government and the Regions, Road Safety Advisory Panel Papers, http://www.roads.dtlr.gov.uk/roadsafety/strategy/rsap/03.htm}

Samuels, M.L. (1993) "Simpson's paradox and related phenomena", Journal of the American Statistical Association, 88(421): 81 - 88

Schieber, R.A. and N.J. Thompson (1996), "Development risk factors for childhood pedestrian injuries", Injury Prevention 2: 228 - 236

Sharples, J. and J. Fletcher (2001), "Tourist Road Accidents in Rural Scotland", Development Department Research Programme Research Findings Report Number 119, Scottish Executive Central Research Unit

Silburn, R., D. Lucas, R. Page and L. Hanna (1999), "Neighbourhood images in Nottingham: Social cohesion and neighbourhood change", Joseph Rowntree Foundation

Stevenson, M., K. Jamrozik and P. Burton (1996), "A case-control study of child pedestrian injuries in Perth, Western Australia", J. Epidemiol. Community Health 50: 280 - 287

Stevenson, M., H. Iredell, P. Howat, D. Cross and M. Hall (1999), "Measuring community / environmental interventions: the Child Pedestrian Injury Prevention Project", Injury Prevention 5: 26 - 30

Thomson, J.A., A. Tolmie and T. Mamoon (2001), "Road accident involvement of children from ethnic minorities: a literature review", DETR, London, Road Safety Research Report No. 19

Towner, E.M.L., S.N.Jarvis, S.S.M.Walsh and A.Aynesley-Green (1994), "Measuring exposure to injury risk in schoolchildren aged 11 - 14", British Medical Journal, 308: 449 - 452

White, D., R. Raeside and D. Barker (2000), "Road accidents and children living in disadvantaged areas: a literature review", Scottish Executive Central Research Unit, Research Findings 81

Wilkins, G., P.Hayward and K.Johnson (2000), "Home Office Statistical Bulletin 26/00 Motoring Offences England and Wales 1998 and 1999", Home Office, London

Paul Hewson
mr.tex@statisticulate.uklinux.net