Prediction of violence in inpatient settings

Abstract

Taking into account the high incidence of violence in psychiatric institutions (see Chapter 3) and the hazards posed thereby to the physical and mental health of staff, there is a need for procedures and instruments for the prediction of violence in clinical routine. This applies not just to safety within institutions but also clinicians who have to deliberate patients' rights and the possible danger to the public before their discharge. In the UK, practice guidelines suggest that best practice should include a documented risk assessment before any patient is discharged to the community (Reed, 1997). However, to develop and evaluate formal instruments is a second step after robust predictors of violence have been determined. At first glance, the conditions seem similar regarding community violence and the risk of violence posed by discharged patients with mental disorders. A large body of research has identified a set of variables which are consistently associated with violent delinquency, among them personality traits, nonspecific and specific historical items, psychiatric diagnoses, and clinical symptoms (Bonta, Law, &Hanson, 1998; Dolan &Doyle, 2000; Harris &Rice, 1997; Hodgins, Mednick, Brennan, Schulsinger, &Engberg, 1996; Swanson, Holzer, Ganju, &Iono, 1990). Thereafter, predictive instruments such as PCL-R (Hare, 2003), the HCR-20 (Douglas, Ogloff, Nicholls, &Grant, 1991), and the ICT (Monahan et al., 2000) have been developed and evaluated within the last decade (Webster, Douglas, Belfrage, &Link, 2000). Repeatedly, it has been shown that static and historical variables make a major contribution to the incidence of violence in the community than clinical variables which can underlie therapeutic influences (Bonta et al.; Buchanan, 1999; Dolan &Doyle; Harris &Rice). Many studies with rather different methodological approaches provide evidence that the risk of violence in the community is increased by a diagnosis of major mental disorder (Eronen, Hakola, &Tiihonen, 1996; Hodgins et al.; Swanson et al., not confirmed, however, by the MacArthur study findings, Steadman et al., 1998), substance abuse (Hodgins et al.; Räsänen et al., 1998; Swanson et al.), ideas of persecution, especially so-called "threat control override" symptoms (Boeker &Haefner, 1973; Link, Stueve, &Phelan, 1998), medication noncompliance (Swartz et al., 1998), "concentrated poverty" in the neighborhood (Silver, Mulvey, &Monahan, 1999), history of violence, male gender, and young age (Bonta et al.; Monahan &Appelbaum, 2000).