Violence in psychiatry, view‐points for standardized research

Violent and aggressive behaviour is a complex phenomenon of great importance in psychiatry. Unfortunately, both de®nitions of aggression as construct as well as de®nitions of its expressions are several and con ̄icting in the research literature. Aggression can be described both by its expressed behaviour and by its inner experienced emotions. These two aspects exemplify different dimensions of aggression as construct (1). In different psychological theories, aggression as phenomenon is regarded as drive, defence mechanism, learned behaviour, personality trait, response to frustrated arousal, etc. The conceptual problems connected with aggression is probably semantic, i.e. different theories use the term aggression to very different, sometimes overlapping, constructs. The conceptual problems surrounding aggression is, however, most probably a consequence of the multidimensional character of these phenomena. In 1987, we suggested a multidimensional approach to different aggression constructs with dichotomous categorization along to three dimensions: 1) inner experience vs. outward behaviour; 2) aggressor's view vs. observer's view and; 3) persistent state vs. episodical occurrence (1). In clinical psychiatric practice and research, for example, aggression is often referred to as aggressive or violent behaviour. A common-sense de®nition, probably common, but not always laid open, in earlier research and clinical practice, is `behaviour that from an observer is regarded as aggressive'. This kind of `common sense'-de®nition has a face validity in relation to the application of research results, but its construct validity would be de®ned by the dimensions as `aggressive behaviour', `observer's view' and `episodical occurrence'. Violence also gives rise to complex reactions among both victims and other more or less related people because of its interactional and emotional dimensions. Emotional reactions following violence tend to interfere with rational analyses of causes of violence, especially when psychiatric patients are involved. Reactions in the public press bear witness to this phenomenon. Emotionally driven solutions are often prematurely presented on the problem of violence from psychiatric patients without any rational analysis (2). This phenomenon is also evident in psychiatric practice. Disruptive acts of violence inside psychiatric treatment facilities often cause emotionally clouded reactions from staff and medical of®cers, sometimes leading to meaningless interventions, possibly more deleterious for both patients and caregivers than no reaction at all. Adequate measures require adequate knowledge. Therefore, it is of the utmost importance that clinicians and caregivers have the possibility to correctly predict and prevent violent acts from patients, based on accurate knowledge and information. In order to perform relevant predictions several pitfalls have to be overcome. In the research literature, many risk factors associated with violent behaviour have been presented. However, signi®cant relations do not imply that a risk factor has a predictive capacity in every setting and time perspective. For example, risk factors of demographic character (e.g. age, sex, previous history of assault), predicting rare events in the distant future do not predict violence within the closer time perspective (3). Even though promising preliminary work has been performed (4), much research is still lacking regarding how to predict violent behaviour in the `ultra-short' time perspective, i.e. how to detect violent behaviour in psychiatric settings within the coming few hours. It should also be presupposed that not only different patient groups, but also different treatment settings have different risk pro®les. Another important problem when trying to evaluate studies on aggression in order to ®nd valid predictors is the lack of consistent methodology in earlier research. In earlier studies, aggression and violence has been described without consistency and without the possibility to compare with other studies. The previous lack of methodology has led to con ̄icting results regarding important predictors of aggressive behaviour. Reviews on aggression prediction have, until recently, been ̄awed by the methodological inconsistencies in earlier studies. During the last decade, however, several valid and reliable instruments for This issue of Acta Psychiatrica Scandinavica also includes short biographies of Tom Palmstierna and BoÈrje Wistedt. Acta Psychiatr Scand 2000: 102: 79±80 Printed in UK. All rights reserved Copyright # Munksgaard 2000