|
|
|
|
40. There would appear to be relatively few - either national or international - authorities in this field. In fact, much of the literature seems opinion based. It is difficult to relate the issues of practice to a body of authoritative research data, which often is descriptive rather than methodologically rigorous and it is disturbing that there is no consensus on a number of fundamental issues.
41. There are related definitional problems, in that the relatively well defined area of dangerousness has been subsumed within the wider framework of risk. The limits of risk can be more difficult to set down. A conflict was noted between the legal view (which tends to consider the issues as categorical) and professional or clinical guidelines (which emphasise the uncertainty and the danger of over-conservatism in evaluation) leading to the risk of additional harm being inflicted. A central theme is the argument as to whether actuarial (population statistics) or clinical methods offer the more reliable predictors.
42. A clinical view might be that such a distinction is artificial, as population data provides the tools to inform clinical decision making; one without the other is less robust. The apparent incompatibility of the clinical and legal views may be explained by the complexity of the phenomena and different populations. For example, an instrument used on a North American prison population may have limited predictive value, say, for a long-stay forensic patient group in Scotland. At best, actuarial tools do have the advantage of comparison with a larger population over a period of time and may help focus clinicians' considerations on all aspects of the situation.
43. Fuller and Cowan27 point to inherent difficulties that narrow the application to strategic decision making and classification tasks. Further the care and management of mentally disordered offenders can involve not only the assessment of violence or criminogenic risk, but also a diversity of risk criteria over the short to medium term, including behaviour harmful to self, to others and property, psychiatric relapse, and risk to the public at large. The authors know of no actuarial risk assessment instrument, extant or in prospect, with the versatility for the task typically facing forensic services.
44. There is an increasing expertise developing in actuarial methods of risk assessment with an apparent divergence from the practicalities of risk management. As various assessment instruments are tested it seems likely that those, such as the HCR-20, that combine actuarial and clinical assessment will have greatest usefulness in terms of the individual patient and his/her care.
45. The difficult interface between risk assessment and risk management is recognised. Lindqvist and Skipworth28 offer that relatively static risk factors such as personal demographics and personality characteristics are uncommon targets for rehabilitation but they acknowledge that they form the core of all risk assessment tools currently in use. The delineation of risk factors between static and dynamic in terms of mental illness may be helpful in terms of planning treatment and risk management.
46. Risk explained in terms of probabilities may give a false veneer of precision in relation to the individual. Obviously, a high risk managed well may be safer than a low risk managed poorly. The risk exists that some falsely associated with a high actuarial risk score (false positives) will be locked up because of "guilt by statistical association". Equally unacceptable is the risk that those with a falsely reassuring low actuarial risk score (false negatives) may be considered ready for release, despite clinical concerns. Finally, the correctly identified group with a high probability of future violence can be in danger of having the "dangerous" label attached to them. The risk here is in their being treated as if they had an immutable quality of viciousness rather than possessing a range of properties and pre-dispositions, open to modifications.
47. No system or administrative measure can compensate for a failure to invest in the resources necessary to provide balanced multi-disciplinary working throughout the services providing care.
48. There does seem to be agreement that a past history of violence is a reliable predictor of future violence. Alcohol makes the related expression of violence more likely for those with severe and/or enduring mental health problems. A past history of violence combined with substance misuse considerably increases the likelihood of violence taking place.
49. The best practical model seems to be a consistent comprehensive clinical evaluation, based on detailed information gathering, by aware and appropriately trained individuals, who keep in mind the wider picture. This is about good practice principles, but there is no robust and consistent expertise or particular skill which can be transferred to practitioners in the field. A review by Mullen16 distinguishes dispositional, historical, contextual and illness related factors and helpfully places these into an ethical framework. He suggests predictions of dangerousness should be:
|
|
|