Courts and review panels of various kinds persist in requiring that criminal justice and mental health authorities make predictions about the possible violent conduct of patients and prisoners. The demand for such assessments seems to be increasing rather than decreasing despite the well published evedence from many studies that clinicians as a whole do not possess the required expertise.
The Violence Prediction Scheme (VPS) starts out by reviewing the current state of affairs. It then proceeds to suggest that overall prediction accuracy can likely be improved substantially if: (a) clinicians disqualify themselves from making forecasts in instances where they lack competence and impartiality, and (b) having agreed to
conducet an evaluation, they set about doing the task in an orderly and systematic fashion.
The guide is grounded in the general psychiatric and psychological literatures. The evaluation is completed in two stages. The first part, actuarial, stresses the use of a Risk Assessment Guide (RAG) to isolate twelve key factors from file and background materials. The variables are then weighted and combined into a form which yields a specific probability figure. Factors and weights are derived from research based on a particular large sample of mentally disordered men who had in the past committed serious violent offences and whose conduct had been followed for some seven years after release from a secure hospital. The second stage, clinical, offers evaluators a set of ten variables which singly or in combination allow the clinician to exercise some discretion in modifying the actuarial probability figure up or down. The clinical items are organized for convenience around a mnemonic device called ASSESS-LIST. Examples of ASSESS-LIST items are antecedent history, self presentation, supervision, and treatment progress. Clinicians are urged to be conservative in making adjustments to the previoulsy-established RAG scores. Stress is laid on the idea that the scheme outlined in the text will require further validation.
Whatever the eventual reliability and accuracy of the VPS may prove to be, close consideration of items in the clinical indicators list should, in the meantime, suggest approaches to treatment and remediation which may lower risk to the public. Certainly it will need modification and substantiation when used for purposes different than those described in this guide. The point is made that those who conduct evaluations have a responsibility to ensure that their predictions are amenable to analysis according to a scheme such as offered here or some other generally agreed and publicly-stated device.
--- from the publisher
Table of Contents:
List of Tables and Figures
Framing the Issues: The Literature on Prediction of Violence
Coping with Demand: Legal and Political Pressures
Deciding to Assess: Approaching or Avoiding the Task
Assessing Clinical Factors: Integrating Clinical and Actuarial Judgements
Conveying the Results: Preparation of Reports
Appendix A: An Example of a Psycho-Social Assessment
Appendix B: An Example of the Actuarial Risk Appraisal
Appendix C: Static Predictors of Violent Recidivism in Mentally Disordered Offenders
Notes on Authors