In the largest randomised trial to date comparing intensive case management with standard care in psychosis, no significant reduction in violence was found. Risk factors for violence included previous violence, drug misuse, younger age, and victimisation, confirming the results of previous studies in psychotic patients.12,13
Violence was also associated with a history of learning difficulties, a factor previously identified in non-psychotic populations.14
Strengths and weaknesses of the study
Different methods for measuring violence can produce very different prevalences. The recent use of multiple combined measures, as in this study, has highlighted the limitations of most previous studies, which relied on a single source. One study that specifically compared the yield of violence when different sources were used revealed a dramatically different picture of violence by patients depending on the source of information used.15
Our results support the observation that self report methods consistently produce a higher frequency of violence than use of other records.16
In a small proportion of cases the WHO life chart was completed from sources other than the patient, so the 94% response rate is a slight overestimate.
The optimal prevalence estimate would have been detected with 100% follow up on all data sources. Although we did not achieve this, we did obtain information on all participants from at least one source. One possible source of bias in this study is that intensive case managers may have detected more violent acts and that standard case managers may have under-reported violence. This could conceal an actual reduction in violence in the intensive group. This is unlikely to be the case as the interviews with case managers added only 15 participants who had not been identified by self report or review of case notes. These cases were evenly distributed between the groups. Additionally, we included only actual assaults, and not threats, in our definition so it is likely that most of these more serious incidents will have been detected irrespective of treatment allocation.
Possible bias arising from interviewers not always being blind to treatment group will have been minimised by the use of multiple data sources. The use of validated questionnaires and continual data monitoring at each centre and centrally will have maximised the robustness of the data. Participants were recruited from inner city locations, and results may not be generalisable to other settings. The multicentre design with over 700 patients should, however, increase the external validity.
Prevalence of violent behaviour
The finding that 22% of patients committed assault over the two year period is of concern but concurs with previous work. Studies indicate that between 10% and 40% of patients commit assault before admission to hospital,and the MacArthur risk assessment studyfound that 27.5% of discharged psychiatric patients committed at least one violent act within a year of discharge.17,18
Our study includes violence by both inpatients and outpatients.
Although intensively case managed patients received more face to face contacts with their case managers and more attention dedicated to medication, engagement, daily living skills, housing, and the criminal justice system, we found no evidence that intensive case management reduced the prevalence of violent behaviour over two years. This finding is not challenged by any of the published trials in this field. There have been at least seven randomised controlled trials examining the efficacy of assertive community treatment—the form of intensive case management favoured in the United States—that have included time in jail or legal contacts as an outcome measure.19–25
None has examined violence specifically, and only two of the seven reported reductions in time in jail.21,22
Differences in the organisation of services, in particular the absence of coordinated care in American standard practice, limit the generalisability of these findings to the British setting.
A randomised trial of the management of care by social services conducted in homeless people with severe mental illness in Oxford found a significant reduction in deviant behaviour in the care management group at 14 months' follow up in comparison with care as usual.26
Although this result was encouraging, the study did not examine violent behaviour specifically, the intensity of the intervention was decided by the individual's care manager, and the level of care received by the control group was unclear. We must therefore conclude that intensive case management, or indeed assertive community treatment, has shown no efficacy in reducing violent behaviour in severely mentally ill patients.
Implications of the study
It remains unclear why intensive community treatment has such a negligible effect on illegal behaviours. In those studies examining time in jail as a secondary outcome the base rate of time spent in jail may have been too low to detect a change in some samples. Alternatively, assertive community treatment and intensive case management have been designed as vehicles for providing clinical services and reducing reliance on inpatient facilities, and these interventions may need considerable modification to address the different needs of patients who are prone to engage in violent or illegal behaviour. Specific interventions to improve compliance with or uptake of treatment for substance misuse are probably important. More controlled research on this question is needed.
Despite the lack of empirical studies on the effect of increasing the intensity of treatment in the community on violence in general psychiatric or forensic populations, research in the United States is now focusing on the effect of combining community treatment with legally enforceable interventions to reduce violence. A recent study, with some important limitations in its methods, found that outpatient commitment (enforced community treatment) for longer than six months combined with regular services resulted in a significant reduction in community violence in severely mentally ill patients at risk of violence. Neither outpatient commitment nor regular services alone was effective.27
Similar legislation for compulsory community treatment in England and Wales has recently been proposed in a government white paper.28
Future research may have the challenging task of evaluating the effectiveness of combining specific clinical interventions within or without a protective legal framework.