%0 Conference Paper %A Nunn, James %A Barnes, Jo %A Morris, Andrew %A Petherick, Emily %A Mackenzie, Roderick %A Staton, Matt %D 2018 %T Identifying MAIS3+ injury severity collisions in UK Police collision records %U https://repository.lboro.ac.uk/articles/conference_contribution/Identifying_MAIS3_injury_severity_collisions_in_UK_Police_collision_records/9340865 %2 https://repository.lboro.ac.uk/ndownloader/files/16949555 %K MAIS3+ %K Data %K Linking %K STATS19 %K Trauma Audit and Research Network (TARN) %K Collision %K Medical and Health Sciences not elsewhere classified %X Objective This study represents the first stage of a project to identify serious injury, at the level of MAIS3+ (excluding fatal collisions) from within the police collision data. The resulting data will then be used to identify the vehicle drivers concerned and in later studies these will be culpability-scored and profiled to allow targeting of interventions. Method UK Police collision data known as STATS19 for the county of Cambridgeshire (UK) was linked using Stata with Trauma Audit and Research Network (TARN) hospital trauma patient data for the same geographical area for the period April 2012 to March 2017. Linking was two-stage; firstly, a deterministic process followed by a probabilistic process. Results The linked records represent an individual trauma patient from TARN data linked to an individual trauma casualty from STATS19 data. Full collision data for the incident resulting in the trauma casualty was extracted. The resulting subset of collisions has the MAIS3+ injury criteria applied. From the 10,498 recorded collisions the deterministic linking process was successful in linking 257 MAIS3+ trauma patients to collision injury subjects from 232 separate collisions with the probabilistic process linking a further 22 MAIS3+ subjects from 21 collision events. The combined collision data for the 253 collisions involved 434 motor vehicle drivers. Conclusions We produced viable results from the available data to identify MAIS3+ collisions from the overall collision data. %I Loughborough University