Kevin S. Douglas

A Prevention-Based Paradigm for Violence Risk Assessment: Clinical and Research Applications

The rationale for this article was to outline and describe an emerging model of prevention-based violence risk assessment and management and to discuss attendant clinical and research implications. This model draws on structured professional judgment rather than on actuarial prediction or unstructured clinical prediction. Its purpose is to prevent violence through the assessment of relevant violence risk factors and the application of risk management and intervention strategies that flow directly from these factors. The authors discuss the nature of the clinical tasks that stem from the model as well as a four-step validation procedure required to evaluate it.

Treatment Effects on Forensic Psychiatric Patients Measured with the HCR-20 Violence Risk Assessment Scheme

This study focuses on the foundational issue of change in violence risk factors in forensic psychiatric patients across multiple assessment periods using the HCR-20 scheme. HCR-20 assessments performed on 150 forensic psychiatric patients at two special (forensic) hospitals in Sweden were studied in relation to the patients time at institution, with the aim to investigate a possible relation between level of risk and time at institution. A sub-group was followed across an 18-month period of institutionalization to investigate possible changes in risk factors over time. The latter patients were assessed with the HCR-20 three times, with six months between the assessments. Patients with longer treatment times had lower scores on the HCR-20 Clinical and Risk Management sub-scales, compared to patients who had been admitted to the hospitals for a shorter period of time. This finding was particularly evident on the cross-sectional analyses with respect to the R scale. The clinical risk factors in the C scale also dropped overall in connection with treatment time across both cross-sectional and prospective analyses, although not as substantially as the R scale. The results are consistent with and provide support for the HCR-20 systems proposition that the C and R scales are dynamic, or prone to change.

Capturing Change: An Approach to Managing Violence and Improving Mental Health

Recently colleagues of ours discoursed on the topic of treatment effectiveness as it relates to persons with serious mental and personality disorder (Harris & Rice, 1997). They introduced their piece by pointing out that, although they had never encountered a fully effective intervention programme, they would in all likelihood recognize one such if they were to see it. Our present outlook is similar to that of Harris and Rice but goes a step further. We suggest that even if a programme is actually effective, chances are that behavioural and attitudinal changes will be indexed so haphazardly, if at all, that they will not be captured and so will not enter the record. And if positive changes are not pinned down during treatment or intervention the result can be to the great detriment of the individual patient, prisoner, or parolee. By the same token, if absence of changes or negative changes are not taken into account, innocent members of society can be placed at risk for violence. It will be argued here that, despite the almost mesmerizing abundance of validated, semi-validated, and unvalidated tests and scales for risk assessment on the market, there exist in fact few clinically relevant, procedurally fair devices for measuring changes in riskrelevant dynamic factors on an interdisciplinary basis. This would apply to measurements both in institutions and in the community. We would also like to posit that serious attempts to document dynamic changes could very possibly have salutary effects, meaning that the mere focus on indexing changes may play an important if minor role in inducing them. The notion of dynamic change requires some comment. Later in the chapter, especially as we discuss the Historical, Clinical Risk (HCR-20) Scheme (Webster, Douglas, Eaves, & Hart, 1997), we draw a distinction between static, more-or-Iess demographic, largely unchangeable, file-based factors and dynamic, alterable, mainly clinical variables. In this chapter emphasis is placed much more on the latter than on the former. We are interested in exploration of variables known to have the potential to elevate or reduce risk of violence in institutions and in the community. If these variables can be agreed upon, defined, and measured, there is at least some

Multiple Facets of Risk for Violence: The Impact of Judgmental Specificity on Structured Decisions About Violence Risk

Among the many advances that have occurred in the field of violence risk assessment over the past decade has been the realization that violence is multifaceted. Violence and the risk for violence include many facets including the severity, density, imminence, target and nature of violence. Despite advances in the understanding and conceptualization of violence and risk, very few empirical risk assessment studies have considered the multifaceted nature of violence. Using a structured professional judgment model of risk assessment, the present study evaluated whether more highly specific judgments of violence (i.e., low, moderate, or high risk for short-term violence, or serious violence, versus violence generally) could be made with reasonable reliability and incremental validity. The findings suggested that more specific judgments about particular facets of violence were not made with the same reliability and validity as either the numerical scores or the omnibus structured clinical risk ratings for general violence. The possible reasons for the findings are discussed with suggestions for future research to continue to study this important area.

Juveniles Evaluated Incompetence to Proceed: Characteristics and Quality of Mental Health Professionals’ Evaluations

The authors studied quality of evaluations of juveniles whose competence to proceed was at issue before the court and characteristics of mental health professionals completing these reports. Many evaluators failed to address important legal issues and offered incomplete descriptions of the youths, their capacities, and the evaluation techniques used. Evaluators commented on legal competency factors in most reports, as required by law, but often failed to identify the cause of the noted incapacity or whether children met commitment criteria (also required by law). Although core competence-related abilities identified in Florida law were addressed in the majority of evaluations, assessments were less than optimal in other psycholegal and clinical domains. Seven recommendations are offered for conducting juvenile competency evaluations that are professionally sound and meet legal requirements.

Evaluation of the Hare P-SCAN in a non-clinical population ☆

Abstract Using 100 university undergraduates as participants, this study examined the structural reliability and construct validity of Hare and Herves P-SCAN (Hare & Herve, 1999) a 90-item measure that purports to serve as an “… early warning system” or “rough screening device” (p. 1) for identifying individuals with substantial psychopathic features. Internal consistency indices (e.g. Cronbachs α, mean inter-item correlations) for the three P-SCAN subscales (Interpersonal, Affective, Lifestyle) suggested excellent reliability. Statistically significant, though modest correlations (range 0.21–0.33) were obtained in 5 of 6 comparisons with the self-report Primary psychopathy and Secondary psychopathy scales developed by Levenson, Kiehl, and Fitzpatrick (1995) for use with non-institutional populations. Evidence for the external validity of the P-SCAN was obtained via significant positive correlations (range 0.22–0.24) with participants scores on a self-report measure of antisocial activity. Limitations of this study and interpretations of P-SCAN as a measure of psychopathic features are discussed.