Enhanced understanding of risk assessment in police custody in England and Wales using statistical modelling

The issue investigated in this paper is the risk assessment of detainees whilst they are booked into police custody in England and Wales. It is estimated that hundreds of thousands of people pass through police custody each year, with each detainee undergoing a risk assessment to establish risk of harm to themselves or to others. However, there has been little research to establish which risk factors identified through the current risk assessment process used within England and Wales have the most impact on risk management. Currently the analysis and evaluation of risk is subjective, based on the custody officer’s judgement and experience. Little is known through literature as to which factors are most influential in this risk assessment process. This paper highlights these gaps in the process and identifies the driving factors in decision making around detainee risk in an attempt to understand the risk management process better before detailing what further research can be conducted to close the gap in the process. Multivariate analysis was carried out on a data-set comprising of custody record data from three police forces, in particular information recorded as part of the risk assessment. Logistic regression, decision tree and discriminant analysis methods were used as all have the ability to classify and predict. Comprehensive analysis and augmentation of results has determined a set of variables which had an influencing impact on observation level. These results provide the background to the development of a supportive and robust risk assessment tool.