Targeting road injury prevention (TRIP)

Organisation
Addenbrooke’s Charitable Trust & Cambridgeshire Road Safety Partnership

Amount awarded
£100,000

Completed
2021

Uploaded to Knowledge Centre
21 September 2021

Summary
In this study, the Cambridgeshire & Peterborough Road Safety Partnership set out to explore new ways that road safety interventions can be delivered to reduce serious and fatal injuries resulting from collisions – by targeting specific drivers who are responsible for these collisions, based on geodemographic profiles.

STATS19 was mapped to a culpability tool for the first time, and the results indicate the potential of using this methodology to identify drivers causing collisions – and to use this knowledge to target specific road safety interventions.

The methods used in the study are now being validated for use on a larger dataset.

In more detail
Prior to 2012 the UK saw a sustained reduction in casualties and deaths as a result of road collisions. However, since then there has been a general plateauing of annual road deaths, and incidents of serious injuries have also followed this trend.

This proof-of-concept study explored the available data and methods involved to enable routine use of geodemographic profiling when developing road safety interventions.

Methods used to identify ‘culpable’ drivers
The following methods were used to identify Cambridgeshire drivers who were culpable of causing these collisions, and to establish whether they were different to non-culpable drivers:

• The police collision database (STATS19) was linked with hospital trauma audit research network (TARN) data for a five-year period to identify Cambridgeshire resident drivers who were involved in clinically defined serious injury collisions.

• All drivers were ‘culpability scored’ and categorised as being culpable, contributory, or non-culpable for the collision. To achieve this the STATS19 variables were mapped on to an existing tool.

• Full geodemographic profiles were appended to the drivers with a culpability score.

• Analysis of the data investigated the culpability and geodemographic profiles of the drivers and explored differences in drivers to inform road safety interventions.

Results
The study identified 564 drivers involved in a serious injury or fatal collision on the Cambridgeshire road network, who had a culpability score. 

The mean age of these drivers was 43 years (SD17) and most were male (434, 77%). For these drivers, the significant factor impacting on the odds of being culpable was their age; being under 26 years or over 76 years showed higher odds compared to the mid-age range (46-55 years). Being a PTW rider also increased the likelihood of culpability, compared to car drivers.

Being the driver of an agricultural vehicle or goods vehicle showed lower odds of being culpable compared to car drivers, as did residing at an address with an Acorn Type designation of 6 (‘financially comfortable families’) compared to the most frequent Type 23 (‘owner occupiers in small villages’). 

When only considering Cambridgeshire-resident drivers and riders (367 (65%)) the significant factors impacting on the likelihood of being culpable were being a PTW rider, or an Index of Multiple Deprivation (IMD) in the 6th decile compared to the most frequent IMD 5th decile.

The use of risk indexation was explored for the geodemographic Types to identify if there were any Types overrepresented in the study sample compared to the overall population of Cambridgeshire.

Overrepresentation on the risk index determines the extent to which a Type is found culpable compared to the general population across the county.

Type 41 culpable drivers (‘labouring semi-rural estates’) were high frequency for fatal collisions and overrepresented compared to the general population (risk index >300). For serious collisions, Type 23 (‘owner occupiers in small villages’) were high frequency and overrepresented and had a risk index >200.

Conversely, some Types were underrepresented on the risk index, specifically Type 10 (‘better-off villagers’) and Type 5 (‘wealthy countryside commuters’), suggesting lower risk of culpability. 

The authors say it would be interesting to explore this further with larger datasets, to understand how typical the over or underrepresentation of culpable drivers is.

Conclusions
Overall, the methods used have allowed culpable drivers causing clinically defined serious injuries to be identified.

STATS19 has been mapped to a culpability tool for the first time and is being validated for use on a larger dataset.

The results indicate the potential of using this methodology to identify drivers causing collisions, and to use this knowledge to target specific road safety interventions.

However, the sample was small and any inferences in the data need to be made with caution as the focus has been on serious and fatal injury collisions and not those with minor or damage only outcomes, and are limited to Cambridgeshire.  

Road safety implications
The method devised in this study would enhance opportunities for road safety professionals to develop targeted and innovative road safety interventions at culpable drivers.

However, the automaticity of determining culpability from STATS19 variables is required before the method can become user friendly in the real world.

It would also be beneficial to explore the nuances of the geodemographic Types identified in the study, with residents from the profile Types. This would identify whether the profile descriptions have any similarities with residents and determine the best method of delivery of road safety interventions. 

This, in turn, would enhance their effectiveness at reaching the target audience and subsequently reducing serious injury and fatal collisions.

For more information and to download the full project report, visit the Road Safety Trust website:

https://www.roadsafetytrust.org.uk/funded-projects/16/addenbrookes-charitable-trust