Transport CommitteeWritten evidence from Road Safety Analysis Limited

Weaknesses in the Key Outcome Indicators

1. This submission is made by Road Safety Analysis Limited (RSA), a British based social enterprise supporting the road safety industry with internationally pioneering approaches to road casualty analysis and intelligence led interventions. RSA are also signatories to a wider submission in association with a range of organisations who collectively are central to the delivery of road safety in Britain.

2. RSA have determined to make one further submission to express their professional view in respect of the Government’s outcomes framework, and its sufficiency for either the fulfilment of their road safety aspirations or indeed their aims to deliver in more locally relevant and accountable ways. RSA published the first analysis of the national outcome indicators in early September 2011, and the only analysis to date of the local outcome indicators and the relative performance of local authorities. RSA have utilised this experience and their years of experience working with road casualty data to assess the sufficiency of the Key Outcome Indicators (KOIs) in particular.

The Focus on Fatalities

3. All national indicators are based on death rate—this is unacceptably limited and may become increasing meaningless if current progress is continued. With road deaths having fallen to 1,850 in 2010 the lack of clearly identifiable trends in the fatality statistics, notwithstanding the improvement introduced through rate based monitoring, will potentially mask other emerging trends that would show up if broader data sets such as “KSI casualties” and “all casualties” were scrutinised. It seems short-sighted only to utilise a 1% subset of the available data to measure performance.

The Right Measure for Rates

4. Having welcomed the increased focus on rates within the KOIs, it is important to ensure that the right measures are used in calculating these rates. In the case of rates calculated per billion vehicle miles, the more appropriate measure would be collisions rather than casualties.

5. In a rate based approach the aim is to examine the frequency of occasions when the system fails for some reason. One system failure is represented by the occurrence of a collision, not by the resultant number of casualties. For instance, in recent years there have been some significant collisions involving coaches full of passengers around junctions with the strategic road network: this type of incident tends to result in large numbers of casualties, and therefore can skew the casualty rate disproportionately, given that only a single failure of the system occurred. This is particularly true for indicators measured at local level, where sample sizes are smaller and therefore more vulnerable to such distortion. It is not that casualties are less important, but the measure needs to deliver what is required; in this case a reliable assessment of failure frequency on the road system.

The Case for Residency

6. The three local indicators that have already been promulgated have the benefit of being easy to understand and communicate; however, they reflect an analytical approach which is dated and fails to empower local citizens. In order to understand road safety issues facing a community, it is vital to examine the risks experienced by its residents on the road network. In some cases a majority of casualties resident in an area are injured in incidents elsewhere, outside the borders of the local authority area where they reside. The dissociation between residency and crash location means that expensive engineering and education initiatives are often focussed only on the road, rather than on those road users who are most likely to be at risk. The analytical tools for examining resident risk have been available for a number of years and would allow communities and local authorities to exercise far greater responsibility for the citizens of their area. Citizens should always lie at the heart of a response, no matter where they are at risk.

7. Risk to residents is best expressed as a population based rate of being a casualty, rather than the collision based approach recommended for traffic based indicators. Resident risk calculations are to some extent subject to data quality concerns particularly over postcode collection, but RSA believes that it is possible to correct existing data adequately to provide a reliable basis for analysis.

(a)Resident risk—all casualties.

(b)Resident risk—all target groups (motorcyclists, cyclists & pedestrians).

(c)Resident risk—child casualties as a proportion of child population.

(d)Resident risk—young drivers involved in collisions (not necessarily casualties) as a proportion of young resident population.

Dealing with Equality

8. The national indicators in many ways are a step change from previous measures of road safety performance. The increased focus on rates is certainly a more sensible approach to setting performance benchmarks, which will be less subject to fluctuation due to exogenous factors such as population migration, the economic cycle or periods of extreme weather; this is to be welcomed. Having said that at a national level there is significant disparity in road risk not just between different types of road user but for road users from different socio-economic backgrounds. It is well knowni that children in more deprived areas are up to four times more likely to be involved in a road traffic collision than those from more affluent areas.

9. Recent research by RSAii has identified that socio-economic factors pertaining to particular community types can correlate highly with particular varieties of increased road risk. This highlights the degree to which those living in deprived and sometimes marginalised communities are further disadvantaged by inequalities in the transport system. RSA would therefore urge the committee to require that the Government considers additional national performance indicators to measure the on-going effort to diminish such inequality with similar indicators calculated at the local level:

(a)National comparative rate (per head of population) between casualties in the lowest IMD decile verses those in the median decile.

(b)National comparative rate (per head of population) between child casualties in the lowest IMD decile verses those in the median decile.

(c)National comparative rate (per head of population) between pedestrian casualties in the lowest IMD decile verses those in the median decile.

(d)Local comparative rate (per head of child population) between resident casualties in the lowest IMD decile verses those in the median decile.

(e)Local comparative rate (per head of child population) between resident child casualties in the lowest IMD decile verses those in the median decile.

(f)Local comparative rate (per head of child population) between resident pedestrian casualties in the lowest IMD decile verses those in the median decile.

Handling Strategic Roads

10. For the purpose of calculating indicators, incidents on the strategic road network should be segregated and reported separately to avoid distortion in measurement of local authority performance. In some areas (West Berkshire and South Gloucestershire are two clear examples), any efforts made to improve road safety performance by the local authority are significantly masked by the presence of large stretches of the strategic road network within their borders for which they bear no direct responsibility; large proportions of collisions and casualties in such areas occur on these roads. In West Berkshire for example 24.8% of casualties (25.1% of KSI casualties)iii occur on roads that are not in the authority’s control; it follows that measurement of the authority’s performance based on figures which include these roads in the analysis is likely to be highly misleading.

References

i Road Accidents and Children Living in Disadvantaged Areas, White et al, 2000.

ii Child Casualties Report, Road Safety Analysis, 2010.

iii Analysis of STAST19 reported road casualties 2006-2010 using MASTOnline
(www.roadsafetyanalysis.org)

October 2011

Prepared 18th July 2012