Acute aggression risk: an early warning signs methodology

Purpose - Predicting the imminence of high risk behaviours in in-patients with schizophrenia is an ongoing concern. The purpose of this paper is to explore the utility, validity and reliability of an adapted early warning signs methodology for dynamic risk assessment. Design/methodology/approach - Nursing staff were interviewed to identify operationally defined early warning signs of high risk behaviours. Frequency of occurrence of the early warning signs and the high risk behaviour were rated over a one week period to establish the predictive validity of the methodology. Findings - Support was found for the reliability of staff ratings of the relevance of identified early warning signs and their occurrence within a specified time period. ROC analysis indicates some modest predictive validity in predicting aggressive risk behaviours but effect sizes were small, and there were high rates of false positive predictions. Research limitations/implications - The small sample size limits generalisability. A longitudinal prospective study to better establish the added predictive power of the method over the use of largely actuarial methods is needed. Originality/value - A dynamic risk assessment methodology to assess changes in risk for inpatients would benefit both staff and inpatients. No such methodology has been assessed to date.