Investigating the dynamic nature of multiple risk assessment instruments in a forensic mental health hospital

Readministration of multiple risk assessment instruments to determine the risk of aggression over the short, medium, and long term is common practice in many forensic mental health settings. Justification for the repeated administration and use of multiple instruments is based on purported differences in discriminative validity of risk factors according to whether they are characterized as static, stable, or acute, and the composition of these tools, in terms of the relative balance of different types of risk factors, which can impact the discriminative validity of these instruments over different follow-up periods. However, research has yet to determine whether the use of multiple and repeated administration of risk assessment instruments improves identification of risk, and it may be the case that the heavy burden on service providers to administer multiple instruments is unwarranted. In this study, time-dependent cox regression with repeated assessments and recurrent events was used to investigate the dynamic nature and incremental validity of the dynamic appraisal of situational aggression (DASA), short-term assessment of risk and treatability (START), and HCR-20^v3, which were repeatedly rated on a sample of forensic mental health inpatients (N = 240) over a 2-year period. Results suggest that using the rolling mean or the most recent risk assessment yielded the most accurate characterization of change in aggression risk. Repeated administration of dynamic risk instrument instruments improved the identification of aggression beyond the initial risk assessment. Although static, stable, and acute factors were significantly related to aggression, the combination of data from multiple risk assessment instruments may not result in clinically meaningful improvements in risk identification. (PsycInfo Database Record (c) 2023 APA, all rights reserved).