What can structured professional judgement tools contribute to management of neurobehavioural disability? Predictive validity of the Short-Term Assessment of Risk and Treatability (START) in acquired brain injury.

Aggression is a frequently cited outcome of acquired brain injury (ABI). Paradoxically, evidence suggests that ABI clinicians underestimate the risk of violence, and aggression in neurobehavioural services appears more frequent than in forensic/secure psychiatric settings. Risk assessment tools are endemic in the latter and highly beneficial in managing harm. However, none has been validated for ABI. This study examines the predictive validity of ABI violence using one established tool, the Short-Term Assessment of Risk and Treatability (START). It is argued that successful management of aggression in neurobehavioural rehabilitation is partly attributable to ongoing programme calibration following regular review of measures conceptualised for ABI. The predictive ability of START and ABI measures was determined through correlational, receiver operating characteristic and hierarchical multiple regression analyses, using 4559 aggression recordings concerning 76 patients over 4 weeks. The START risk of violence was classed "low" for 50% of the sample and "high" for 13.7%. Significant relationships between individual measures and aggression were evident. However, multivariate analysis confirmed that the best fitting model comprised ABI measures of neurobehavioural disability and functional abilities: START data did not increase explained variance. ABI risk prediction may benefit from development of specific tools. However, these should be conceptualised for ABI to incorporate the diverse range of variables underpinning neurobehavioural disability.