Assessing Violence Risk in Psychiatric Inpatients: Useful Tools

The HCR-20 is a clinical risk assessment tool born out of the need to maximize the benefits of using the latest research in predicting violence with the experience of forensic practitioners who routinely assess risks in clinical settings. The goal of the instrument was to lay an effective foundation for completing risk assessments while keeping in mind time pressures and pragmatic obstacles faced by clinicians in clinical settings.

The instrument is divided into 3 sections.

• Historical: includes a review of the person's index offense; criminal, psychiatric, family, school, and vocational background; and other relevant historical information.
• Clinical: entails a description of the person's state of mind at the time of the incident or offense, as well as current diagnostic symptoms and other clinical factors that bear on risk.
• Risk management: suggests a treatment plan and gives indication of what services and supports must be available to the individual if risk or violence is to be managed within acceptable bounds.

Violence risk assessment is based on a manageable number of test items rooted in evidence-based knowledge and organized around a few important cross-disciplinary ideas. The items are defined precisely enough for testing but written so as to invite efficient application to a variety of issues and settings that contain a high proportion of persons with histories of violence and a strong suggestion of mental illness or personality disorder (eg, forensic psychology or psychiatry, parole, or correctional environments). The summary stipulates time periods for which the assessment results are intended to hold, how the prediction might vary with altered situational circumstances, and base rates of violence in pertinent samples.

The HCR-20 consists of a checklist of 20 items: 10 historical factors, which are weighted as heavily as the 5 present clinical variables, and 5 future risk management issues (Table 1). Of the individual items, substance abuse and psychopathy are most strongly correlated with violence. Violence includes verbal aggression, self-directed aggression, and aggression toward others and objects.

Reports and notes by psychologists, psychiatrists, social workers, police, prosecutors, nurses, and other professionals provide the basis for interviewing and testing the patient. The interview not only garners new information but assesses for inconsistencies between what is known about the person and what the person would have the assessor believe. For example, uncorroborated self-reports, when self-incriminating, may be more relevant than uncorroborated self-promoting self-reports. In the criminal and forensic setting, many people attempt to minimize the extent to which they may have harmed others, are responsible for negative actions, or have behaved in harmful ways in the past.

In most cases, file review, interview, and testing are enough to complete the HCR-20. However, for future risk management, consultations with colleagues responsible for treatment or community release plans will likely be needed. For example, reports from previous case managers, social workers, or probation or parole officers are needed to determine how well the individual being assessed fared on past releases. Victim or other collateral interviews may also be helpful. All sources—consulted or not consulted (with explanation)—should be noted in the assessment.

After scoring individual items, a total score is obtained. An individual is then determined (based on the available information) to be at low, moderate, or high risk for violence within parameters of a given timeframe and in a given setting. Although total score bears some relationship to the final risk assessment, the decision is not simply based on a cutoff score. The relationship between the number of risk factors present and risk for violence is probably distinctly nonlinear; risk likely depends on the specific combination, not just the number, of risk factors present. When items are omitted, assessors must qualify their opinions accordingly, acknowledging whether and how their opinions might change if full information were available (Table 2).

The HCR-20 has been tested in a variety of settings: civil psychiatric, forensic, and correctional. An annotated bibliography for the HCR-20 is available at http://www.sfu.ca/psyc/faculty/hart/HCR-20%20Annotated%20Bibliograph,%202006.pdf; it lists more than 50 studies that used the HCR-20 to assess violence risk. Receiver operating curve tests for these studies mostly demonstrate areas under the curve in the 0.7 to 0.8 range, superior to unstructured clinical assessments.

The HCR-20 may be purchased from a source such as Psychological Assessment Resources, Inc. (www3.parinc.com). An introductory kit consisting of the professional manual, 50 coding sheets, and the HCR-20 Violence Risk Management Companion Guide costs $115. Additional coding sheets cost $44 per pad of 50, although they are not technically necessary to complete and document an assessment using the tool.

With the widespread use of well-researched and validated risk assessment tools in current practice, psychiatrists need to be familiar with the strengths and weaknesses of each and arrive at an informed decision about when their use is indicated. Once considered the domain of forensic psychiatrists, violence risk assessment and management research is now informing and guiding the practice of general inpatient psychiatrists throughout the nation.