Detecting and predicting self-harm behaviour in prisoners: a prospective psychometric analysis of three instruments

Abstract

Research has revealed high levels of suicide and self-harm within young adult prisoners, but many studies have not considered the applicability and validity of its measurement for both male and female prisoners. Previous studies have focused on retrospective evaluations of instruments which are not useful evidence in informing clinical practice and decision making. To evaluate the validation and prediction of suicide and self-harm risk in young adult prisoners. The study was divided into two stages. Stage one used a cross-sectional design of 1,166 prisoners across six HM Prisons to validate the use of three questionnaires: the Beck Depression Inventory, the Beck Hopelessness Scale and a newly devised instrument (SCOPE tool). The second stage assessed the predictive validity of the three instruments using a 4-year-follow-up study of female prisoners across two HM Prisons in UK. Self-report and official records were used to measure suicide and self-harm risk. Logistic regression methodology, receiver operator characteristic curves and Youden's index were used to determine the range of thresholds for the three tools. Self-report measurement of suicide and self-harm behaviour using the three instruments presented a range of sensitivity and specificity values (65.9-72.3% and 64.9-74.0%, respectively). Predictive measurement of suicide and self-harm behaviour in the follow-up study presented a range of sensitivity and specificity values (54.6-80% and 62.2-69.4%, respectively). Screening for self-harm and suicidal behaviour in young prisoners has generated a range of cut off points for the identification of those at risk. These serve as a bench mark for service planners and practitioners.