D ynamic A ppraisal of S ituational A ggression lacks utility

In the abstract of a recent article, Griffith et al. (2013) claim that the Dynamic Appraisal of Situational Aggression (DASA) ‘presents as a valid measure for appraising risk of imminent aggression in mainstream mental health inpatient settings’. They base this assertion on the finding that actuarial judgments of future violence in the next 24 hours were more likely to prove correct than clinical judgments. However, the authors fail to report the proportion of those classified as a high risk using the DASA who went on to become violent in the next 24 hours. This proportion is usually referred to as the positive predictive value (PPV) of a risk assessment (Large et al. 2011). The PPV of the DASA in this study can be calculated to be 0.49% using the reported base rate of violence over 24 hours of 0.215%, the reported sensitivity of 68%, and the reported specificity of 70%. This means that for every 204 people categorized as being potentially aggressive in 24 hours according to the DASA, only one person will commit an act of violence. Clearly, any violence-reducing clinical intervention instigated on the basis of a DASA needs to be benign and should not impinge on patient liberties. This is tacitly acknowledged by the authors who suggest that high-risk patients should be ‘distracted or soothed’ by ‘reducing environmental stimuli, engaging the patient in an enjoyable activity such as art, taking a walk, or listening to music’. However, the DASA has a sensitivity of only 68%, meaning that approximately a third of all violent acts are committed by ‘low risk’ patients. It escapes us why people hospitalized with a mental illness who are considered to be at ‘low risk’ should be deprived of soothing and enjoyable activities? Surely, rather than expend precious nursing resources in a futile attempt to predict aggression, we should instead be trying to deliver quality interventions to all patients in mental health units.