Driving while impaired by alcohol: An analysis of drink-drivers involved in UK collisions

Drink-driving represents a critical issue on international organisations’ agenda as one of the key behavioural risk factors in road traffic safety alongside speed, non-use of motorcycle helmets, seat-belts and child restraints. Changing road user behavior regarding these five factors is a critical component in reducing road traffic injuries and casualties. The objective of this study is the identification of drivers who are more likely to contribute to crashes in the UK, while being impaired by alcohol, for designing targeted drink drive compliance campaigns.

Method
To profile drivers with the ‘impaired by alcohol’ factor assigned in collisions, an extensive dataset was used, comprising all reported injury collisions between 2011 and 2015 in the UK (police records), merged with the Experian Mosaic database. A multilevel mixed-effects logistic regression is conducted, utilizing the hierarchical nature of the data (drivers within Mosaic Types).

Results
Using multilevel mixed-effects logistic regression analysis, the finding is that some driver profiles are more likely to contribute to crashes and are assigned the ‘impaired by alcohol’ contributory factor. Drink-related crashes are more common in some circumstances or for some driver groups rather than others. For instance, drivers using single carriageways are significantly more likely to contribute to a drink-drive related crash, as are male or 25 to 35 years old drivers.

Drink-drive related crashes were found to be strongly associated with darkness light conditions and more specific with late night hours (the interval between 03.00 and 04.00 a.m. accounting for a third of the drink-drive related collisions).

Using the Experian Mosaic database and dividing the UK population into 66 ‘Types’ based on demographic, lifestyle, and behaviour characteristics, the finding is that some Mosaic Types are significantly more likely (e.g. Pocket Pensions, Dependent Greys, Streetwise Singles) and other are significantly less likely to contribute to a drink-related crash (e.g. Crowded Kaleidoscope, Cultural Comfort, Penthouse Chic).

Conclusions
The outcome is a more nuanced understanding of drivers contributing to drink-related crashes in the UK. The study concludes by discussing the implications for governments and other co-interested bodies for better targeting and delivery of public education campaigns and interventions.