1. LEPH Conference 2018 James Nunn
Linking police collision data and hospital trauma patient data. Enabling comparison of culpable drivers from serious injury non-fatal collisions with those who cause fatal injuries.
LEPH Conference 2018
James Nunn
Loughborough University Design School, Transport Research Group

2. Background
There had been steady decline in fatal casualties until Since then the numbers have plateaued.
Fatalities. Reported road casualties in Great Britain: 2017 annual report, Department for Transport, 2018

3. Background
The number of casualties with serious injuries has also stopped declining.
New methods of intervention or how to apply interventions may be required to induce a new downwards trend.
Serious injuries. Reported road casualties in Great Britain: 2017 annual report, Department for Transport, 2018

4. Aim
Identification of the people causing serious (MAIS3+) injuries in road traffic collisions.
(AIS - Abbreviated Injury Scale, MAIS – Maximum AIS for a casualty. International system of trauma grading administered by the Association for the Advancement of Automotive Medicine)
Most interventions are based on population stats and do not capture the perpetrators.
Police injury severity classification does not use AIS and is for the serious category is subjective and broad – quadriplegic to a broken little finger
MAIS3+ is the standard adopted by the European Commission to allow international comparison

5. Study 1 Project objective
Link police collision data and hospital trauma patient data.
Identify the serious injury (MAIS3+) subset.
Identify all motor vehicle drivers involved in serious injury (MAIS3+) collisions.
First of 3 Studies
MAIS3+ is the European standard for severe/serious injury.

6. Methods - Data Sets for Linking
Cambridgeshire Police collisions data (STATS19)
No common identifier
East of England Trauma Audit and Research Network Data (TARN)
Four common variables (Age, sex, incident date, first string of postcode) – formatted – no missing data
collisions
23741 subjects
1907 individual trauma patients
STATS19 Department for Transport – Police recorded variables – STATS20
Fatal 158 Collisions – 174 Casualties; 1593 serious – 1823 casualties; 8745 slight – casualties
East of England six counties; Bedfordshire, Hertfordshire, Essex, Cambridgeshire, Suffolk and Norfolk
TARN 43 variables
TARN does not contain all MAIS3+ injuries for East of England – Entry Criteria 72 hrs inpatient – ITU/HDU any stay
TARN already filtered for Collisions

7. Validated Results (MAIS3+)
Methods and Results - Data Linking
1. Deterministic linkage (all four variables must match)
2. Probabilistic linkage (allows variation in age ± 5 years)
Deterministically linked records removed
Full collision data retrieved
Results
295 linked records
257 MAIS3+ linked records
Results
31 linked records
22 MAIS3+ linked records
Validated Results (MAIS3+)
268 linked records
243 collisions
416 motor vehicle drivers
Validated 95.8% true- matches
All records retrieved
Deterministically linked randomly selected 10% (n=26) of the MAIS3+ records
Probabilistically all records
95.8% of the deterministically linked records were true matches.
58.6% of the probabilistically linked records were true matches
Validated 58.6% true- matches

8. Study 2 Project objective
Undertake culpability scoring on the motor vehicle drivers from the serious injury (MAIS3+) subset.
Identify all culpable motor vehicle drivers involved in serious injury (MAIS3+) collisions.
First of 3 Studies
MAIS3+ is the European standard for severe/serious injury.

9. Methods and Results – Culpability Scoring
STATS19 full collision data
Map STATS19 data to Robertson and Drummer (1994) criteria
Robertson and Drummer (1994) culpability scoring tool
Results
263 (63.2%) culpable drivers
31 (7.5%) contributory drivers
118 (28.4%) non-culpable drivers
4 (1.0%) could not be scored
Apply scoring tool criteria to STATS19 data for validated matches to score individual motor vehicle drivers
Robertson and Drummer 1994 chosen as it has a proven track record in the literature and is less subjective than previous scoring tools such a Terhune 1983
Explain principle – Culpable until mitigation applied, if sufficient mitigation achieved, culpability removed
Mapping examples -

10. Next stage
Culpability score using alternative tool (Brubacher, Chan and Asbridge (2012) for comparison Culpability score motor vehicle drivers from fatal collisions for comparison Study 3: Geodemographic profile the culpable drivers from MAIS3+ collisions using their home post codes. Target interventions.
Initial analysis – the culpability is comparable between the two scoring tools, some minor differences relating primarily to single vehicle/ pedestrian collisions
Fatal collisions showed a slightly lowere level of culpability

11. Email: J.Nunn@lboro.ac.uk
Any Questions
Acknowledgements
Co-authors
Jo Barnes, Andrew Morris, Emily Petherick. Loughborough University, Epinal Way, Loughborough, LE11 3TU, UK
Roderick Mackenzie. Cambridge University Hospitals, Addenbrookes Hospital, Hills Road, Cambridge, CB2 0QQ, UK.
Matt Staton. Cambridgeshire County Council, Shire Hall, Castle Hill, Cambridge, CB3 0AP, UK
Thanks to Cambridgeshire County Council, Cambridgeshire Constabulary and Cambridge University Hospital for providing the data and facilitating the process and the Road Safety Trust for funding the project