Authors' reply

We agree with Wand & Large that there currently is very limited support for the use of structured risk assessment instruments as a method for violence prevention. So far only a small number of studies, four including our own, examined this issue. It is troubling that most research efforts seem to focus on the development of new risk assessment instruments and establishing their psychometric properties, rather than on testing the effectiveness of existing instruments. Although identification of predictors and development of instruments are crucial steps in the maturation of both risk assessment and forensic psychiatry, the field needs to move beyond these issues.

The most important risk and protective factors associated with recidivism have by now been established and are agreed on by the research community. There is no disputing the existence of correlations between mental illness, substance misuse, client well-being, quality of life and recidivism. That is why all, or a considerable selection of these factors, are commonly included in risk assessment instruments.^1-3 It seems it is time to move forward and start investigating the benefits of risk assessment instruments and their contribution to more effective treatment interventions in terms of reduction of criminal and violent behaviour. As we ourselves have experienced, introducing randomised trials in clinical practice is difficult, but it can be done, and is an essential step before implementation can be advocated.

A definitive answer about the contribution of structured risk assessment to violence prevention cannot be given at this time. The first signs are not good. The four available studies find either no significant reduction of violent outcome, or the interpretation of their findings is problematic due to differences between study groups at baseline. Differences in clinical setting of the various studies further complicate the integration of findings. Our own data were collected in a community-based forensic mental health setting. In contrast, the other three studies were completed in acute psychiatric (admission) wards. These two settings service different populations, making comparisons less straightforward. It is too early for a proper systematic review on this subject, but the overall picture is not yet convincingly in favour of changing treatment policies by systematically employing structured risk assessment in clinical care.

On the other hand, our paper also shows that proper implementation in clinical care depends on personnel and organisational factors that need to be addressed in a coherent and persistent way before meaningful results can be obtained. The implementation of a randomised controlled trial has its particular challenges, but so does changing clinical practice in and of itself.⁴ As researchers, we may sometimes underestimate this gap between scientific evidence and the changes necessary in clinical practice for the implementation of evidence-based interventions. In order to reach the ultimate goal of prevention and to determine whether structured risk assessment may contribute, more studies are needed that assess the results of properly implemented and already established instruments in different forms of forensic care.