Christopher D. Webster

Violence risk assessment: Science and practice

Within the conceptual framework of the scientist-practitioner model, this paper describes how the science of risk assessment has progressed to a point where it can be of definite assistance to clinicians. Similarly, clinicians can be of marked help to researchers as they play a role in concept definition and as they design and carry out studies to determine the accuracy of their predictions. Risk assessments ought to be carried out in accord with state-of-the-discipline knowledge. Until fairly recently, it may have been argued that the state of knowledge did not provide any sort of reliable or trustworthy direction on violence risk assessment. This position seems no longer tenable. Topic areas discussed include violence and risk assessment generally, the actuarial-clinical prediction debate, the validity of violence risk assessments, predictors of violence, violence risk assessment schemes (empirically validated structured clinical decision-making), communication of risk assessment findings, and implications for training.

The HCR-20 Violence Risk Assessment Scheme - Concurrent Validity in a Sample of Incarcerated Offenders

The Historical, Clinical, and Risk Management (HCR-20) violence risk assessment scheme was coded in a sample of 75 Canadian male, federally sentenced, maximum-security offenders. The concurrent validity of the HCR-20 was assessed through comparison to other risk instruments and to the presence of several past indexes of violent and antisocial behavior. The HCR-20 showed moderate to strong relationships with the concurrent measures. The HCR-20 was as or more strongly related to past violence than were the Psychopathy Checklist—Revised or the Violence Risk Appraisal Guide. Scores above the median of the HCR-20 increased the odds of the presence of various measures of past violence and antisocial behavior by an average of four times. Although recognizing the limitations of a small sample and retrospective design, the results give some indication that the HCR-20 may be worth investigating as a useful tool for violence risk assessments in correctional samples.

The Short-Term Assessment of Risk and Treatability (START): a prospective validation study in a forensic psychiatric sample

A new assessment scheme-the Short-Term Assessment of Risk and Treatability (START)- presents a workable method for assessing risks to self and others encountered in mentally and personality disordered clients. This study aimed to demonstrate (a) prevalence and severity of risk behaviors measured by the START, (b) psychometric properties of START, (c) similarities and differences in START scores across different mental health professionals, and (d) concurrent validity of START with diverse negative outcomes. Treatment team members completed the 20-item, dynamically focused START for 137 forensic psychiatric inpatients. Prevalence and severity of START risk domains were measured for 51 patients detained in the hospital for 1 year. Results revealed high rates of generally low-level adverse events. With some exceptions, START scores were meaningfully associated with outcomes measured by a modified Overt Aggression Scale.

The dimensions of dangerousness

Treatment of dangerousness in both sociolegal research and clinical practice has neglected to consider the multidimensional nature of the construct. An attempt was made to develop a psychometric instrument sensitive to several facets of dangerous behavior among forensic patients. Two trained nonclinical raters used the scale in assessing the dangerousness of 210 patients interviewed in a pretrial forensic clinic. Reliability, item and factor analysis refined the instrument into a 15-item Dangerous Behavior Rating Scheme. Profiles of patients during a two-year follow-up were constructed from contacts with the correctional system and five psychiatric hospitals, and rated on dangerousness to others by nine independent judges. A Pearson product-moment correlation of +0.34 was yielded between aggregated factor scores derived from the instrument and follow-up dangerous behavior. Given the limited ability of either clinicians or psychometric instruments to predict dangerous behavior, and given the socially constructed nature of the dangerousness phenomenon, predictions about future violence should be adopted into practice only under restricted conditions. A number of these limitations are summarized.

Predicting Violence in Mentally and Personality Disordered Individuals

An inmate who has spent the last five years incarcerated for aggravated sexual assault comes before a parole board asking to be considered for release. A person suffering from schizophrenia murdered his parents and was found not guilty by reason of insanity and now, some years later, stands before a release review board. A 14-year-old girl allegedly has attacked a classmate with a knife and a prosecutor now petitions to raise her to adult court. A police officer escorts a young man acting in a threatening manner to a psychiatric emergency service. In each of these instances an important decision must be made that pits the right of the individual not to be arbitrarily detained or punished against the rights of members of society to be safe from potentially violent persons.

A note on portraying the accuracy of violence predictions

We discuss two major reviews of the accuracy of dichotomous (yes/no) violence predictions (Monahan, 1981; Otto, 1992), which reported rates offalse positive errors that were calculated using conceptually and mathematically dissimilar methods. We outline potential problems that result from analyzing data in the form of 2×2 contingency tables and offer some recommendations for future research on the prediction of violence.

Historical-Clinical-Risk Management-20, Version 3 (HCR-20V3): Development and Overview

The HCR-20 Version 3 (HCR-20V3) was published in 2013, after several years of development and revision work. It replaces Version 2, published in 1997, on which there have been more than 200 disseminations based on more than 33,000 cases across 25 countries. This article explains (1) why a revision was necessary, (2) the steps we took in the revision process, (3) key changes between Version 2 and Version 3, and (4) an overview of HCR-20V3s risk factors and administration steps. Recommendations for evaluating Version 3 are provided.

Domains of the impulsivity construct: A factor analytic investigation

Abstract The impulsivity scale from the Personality Research Form (PRF), the control/impulsiveness scale from the Multidimensional Personality questionnaire (MPQ), and the restraint scale from the Guilford-Zimmerman Temperament Survey (GZTS) were each subjected to exploratory factor analysis in a sample of 252 normal adults. The replicability of the derived factor structure for each scale was tested in a sample of 230 undergraduates using confirmatory factor analysis. The relationships among impulsivity dimensions were also examined by subjecting the extracted factors from the three scales to a second order factor analysis. Although developed as unideminsional measures of the impulsivity construct, the MPQ and GZTS scales were found to have a similar multidimensional structure: a cautious versus spontaneous dimension, and a methodical versus disorganized dimension. This multidimensional model is compared with some recently proposed models for the impulsivity construct. The PRF scale was found to be a unidimensional measure that taps only a cautious/spontaneous facet of the impulsivity construct.

A Comparison of General Adult and Forensic Patients with Schizophrenia Living in the Community

Schizophrenia is associated with an elevated risk for violence. The response has been to incarcerate people with schizophrenia and to increase the number of forensic beds. Most of these beds are filled by men with schizophrenic disorders with long histories of offending and of treatment in general psychiatry. Outcome from forensic, as compared to general psychiatric services, is unknown. The present study compared outcome defined as levels of positive and negative symptoms, readmission, and aggressive behavior for 248 men with schizophrenic disorders (150 discharged from forensic hospitals and 98 from general adult wards) during a two-year period after discharge from forensic and general psychiatric services. Patients were intensively assessed at discharge and four times during the subsequent two years. Illicit drug use was assessed both objectively and by self-report. More of the forensic than the general patients had failed to complete high school, had displayed a stable pattern of antisocial behavior since at least mid-adolescence, and had convictions for non-violent and violent offences. At discharge and throughout the follow-up period, general patients displayed higher levels of positive and negative symptoms than forensic patients, and more of them engaged in aggressive behavior towards others. Aggressive behavior was associated with positive symptoms and Antisocial Personality Disorder. Rates of readmission were similar for the two groups. The forensic approach that includes assessing and managing the risk of violence as well as treating symptoms of schizophrenia led to better outcome than that of general psychiatry.

A Review of Arrests Among Psychiatric Patients

E. Glenn Schellenberg,* Donald Wasylenki,** Christopher D. Webster,*** and Paul Goering**** Common wisdom has it that “trouble with the law” is a serious problem for psychiatric patients. This may be especially so for chronically mentally ill patients, most of whom have a diagnosis of schizophrenia. However, little is known about how often clinicians should expect to encounter this problem, or about the factors involved. In a major review of the literature, Rabkin (1979) divided relevant studies into two groups-those published before 1965 and those published after. The earlier studies showed that arrest rates for patients were lower than rates for the general population (Ashley, 1922; Brill & Malzberg, 1962; Cohen & Free- man, 1945; Pollock, 1938). Studies published between 1965 and 1979 showed arrest rates of patients to be equal to, or greater than, those of the general population (Durbin, Pasewark, & Albers, 1977; Giovanni & Gurel, 1967; Rap- peport & Lassen, 1965, 1966; Sosowski, 1974, 1978; Steadman, Cocozza, & Melick, 1978; Zitrin, Hardesty, & Burdock, 1976). Rabkin (1979) attributed the increase in arrest rates to an increase in the proportion of patients with prior criminal records. It is important to note, however, that thousands of patients were “deinstitutionalized” between 1965 and 1974. The bed capacity of U.S. state hospitals diminished from 550,000 in 1961 to 110,000 by 1975 (Bassack & Gerson, 1978), placing many more patients at risk for arrest. Data concerning schizophrenia and risk for arrest were equivocal. Rabkin concluded that schizophrenics are no more likely to be arrested than are other diagnostic groups, with the exception of violent crime, for which schizophren- ics may have higher rates of arrest. She also concluded that psychiatric pa- tients, in general, are not predisposed to be arrested for some crimes as op- posed to others, with the exception of assaultive behavior, for which they are more likely than other persons to be arrested. No conclusions could be reached concerning the effect of psychiatric hospitalization or treatment on arrest rates. This report examines research conducted since 1980 concerned with arrest