John Hinkley has been found not dangerous and is allowed over night family visits. Following his final release from jail John Thanos brutally murdered 3 teenagers in two separate incidents. Every time I hear a story like this one I feel anguish in the pit of my stomach: Columbine, Jonesboro, Littleton, the Morrow Federal Building, the Unibomber, and many more. This doesn’t have to be. Where along the way could these sequences have been changed?
Psychologists and psychiatrists are routinely asked by the Courts and Hospitals, “Doctor, is this person at risk for danger to himself or others?” We know that risk of violence should be assessed separately from mental health issues. Consequently, traditional psychological tests such as the MMPI- 2 (A), MACI, and MCMI-III are not intended to be risk tools and should not be used in that way. We also know from a decade of research the clinical judgment of future risk of violence based on interview alone is little better than chance. Fortunately, the technology to identify those at risk for future violence is developing rapidly. Research on risk evaluation tools has determined that we can improve the prediction of future violence significantly over clinical judgment by using tools developed from statistical analysis of the traits of violent offenders.
Additionally, previous thinking has been that violence is not treatable. However, recent studies indicate that patterns of thinking and behaving can be substantially changed, thus reducing a person’s risk of committing a violent act in the future. Once the level of risk is known, the level of supervision and type of treatment can also be established. Rice, Harris and Quinsey (2002) stated, “Risk assessments using empirically validated instruments can enhance public safety without increasing the number of individuals detained.”
The Additive and Interactive Nature of Violence Risk and Resiliency Factors
When the various social, psychological, biological, and environmental factors have negative family, school, or community components, difficulty managing emotions, lack of empathy, poor self-image, lower levels of problem solving skill, and deficient social skills can be the result. The various social and skill factors that influence children and adults are interactive. There is a threshold of higher numbers of negative factors (abuse, neglect, poor school performance, crime/delinquency) and low levels of positive factors (prosocial activities and friends, and positive future goals) beyond which there is a greater risk for violent behaviors to occur.
Risk Assessment Tools for Youth Violence and Severe Behavior Problems
The completion of the risk assessment tool first requires that the clinician gather a complete psychosocial history, based on record review, direct interviews with the child or adolescent, and interviews with collateral informants such as parents, teachers, therapists, social service agency workers, etc. (American Academy of Child and Adolescent Psychiatry; Rich, 2003). Each risk tool has characteristics that are unique to that tool. By comparing tools, a practitioner can determine which tool(s) are best suited to their needs and their population(s). The tools listed here are not exhaustive, but include those most commonly used by practitioners.
The SAVRY (Borum, Bartel, Forth, 2002) and the EARL-20B and Earl 21G (Augimeri, Webster, Koegl, Levene, 2001) are empirically-based, structured tools for guided clinical assessment. That is, the factors included in the tools are based on pertinent literature in the field, including published studies, and the instruments provide defined factors to be addressed and a specific structure to be followed by the clinician completing the evaluation (Rich, 2003). The PCL-YV (Forth, Kosson, Hare, 1996) and the YLS-CMI (Hogue and Andrews, 1996) have empirically based scoring systems and cut off scores. The DVI (Behavior Data Systems) is self-report and includes suggestions for treatment as does the YLS/CMI.
The CARE (Seifert, 2003) is an easy to use tool for assessing the risk of youth violence and creating a multifaceted case management plan. More than 900 youth with ethnically diverse backgrounds are in the CARE sample. Their ages range from 2 through 19 years, over half had a history of assaults. The participants were taken from all types of settings such as residential, outpatient, and detention centers, a prison for young, violent offenders, as well as groups of youth and adults with mild or no problems. This sample group of clients was from a large geographical area of the United States East Coast and the Mid-West. The total CARE score was significantly correlated with the severity past of behavior problems. Significantly higher CARE scores were seen in those with assaultive histories and these youth were more likely to commit an assault within the next six months.
The CARE assesses both risk and protective factors. The significance of the development of a tool such as the CARE is widespread and offers an opportunity for school, social service, criminal justice, and mental health professionals to access a reliable tool for determining the need for additional testing and for specific treatment protocols. As with the other youth risk tools, the CARE is based on the idea that the more risk factors that an offender has, the greater his risk for recidivism. No one factor predicts youth violence. Each additional factor increases the risk that a youth will be violent. The strongest predictors of past youth violence are: severity of past behavior problems, assault of an authority figure, chronicity of past assaults, psychosis or self-harm, and inappropriate discipline by a caregiver.
Adult Violence Risk Assessment Tools
There have been four stages, thus far in the development of risk evaluation tools. The first generation of violence risk assessments encompassed primarily clinical judgment supplemented by traditional psychological tests, such as the MMPI. However, the MMPI was not intended to measure violence and should not be used in that way. Studies of the accuracy of mental health professionals using unaided clinical judgment to estimate the risk of future violence have demonstrated that psychologists, psychiatrists, social workers, and counselors make inaccurate predictions 80% of the time.
The second generation of tools used clinically derived items that were subjected to statistical analysis. One such tool is the Psychopathic Checklist (PCL). The Psychopathic Checklist (Hare, 1995) is a clinically oriented tool, based on a theoretical construct, Psychopathy. The items center on the behaviors and personality characteristics associated with psychopaths. During research, it was determined that the PCL was also significantly correlated with violent offending and re-offending (r = .34 with past aggression, r = .27 with future violence). There have been many studies of the PCL and it is widely accepted as the “gold standard” in assessing psychopathy, but not necessarily violence.
The Level of Service Inventory (LSI) is also a clinically oriented tool used to determine the level of supervision that an offender is likely to need. It is a combination of clinical items and statistical analysis. The LSI is correlated .27 with new arrests and .43 with general recidivism. These are relatively weak to moderate correlations.
The third generation of actuarial tools includes guided clinical judgment. These tools are more clinically based and usually do not have formal scoring systems. There are some reliability and validity statistics that can be found on these tools (HCR-20 and VRAG). These tools are a combination of clinical items and statistical analysis. The correlation between the HCR-20 and past aggression is .44 to .52. The HCR has a guide to determine risk management procedures.
The fourth generation of tools uses clinical items, resiliency factors, and risk management plans. Dynamic risk factors can measure changes in skills that may help reduce recidivism and static factors are items that do not change, such as an abuse history. These tools include the Adult RME (Seifert, unpublished manuscript).
The RME Risk Management Evaluation) is a violence risk evaluation for adults. It was based on the CARE. The sample includes 250 males and females in an outpatient mental health setting. Like the CARE, it uses risk and resiliency factors and has a risk management planning tool. Studies have found that the risk factors for adult violence are similar or the same as the risk factors for youth violence (Rice, Harris, Quinsey, 2002). Consequently, the RME uses the same items as the youth version, but after further research and analysis the items have been re-worded and item weights have been changed to be appropriate for the adult sample. Both static and dynamic factors, which the research literature identified as being associated with severe behavior problems and aggression were used (Seifert, 2000; Quinsey, Harris, Rice, & Comier, 1998). Higher rates of violent recidivism and other offending behaviors are found when there are greater numbers of risk factors and fewer resiliency factors.
Risk categories for the RME include 1) individual characteristics such as history of violence, poor anger management, psychosis, harming animals and enuresis; 2) peer interactions like bullying behaviors, and deviant peer group; 3) work, school, and educational problems, such as lack of work success; 4) family characteristics, such as exposure to violence during childhood, and a history of harsh disciplinary practices. School and or work success are examples of resiliency factors.
Discussion and Conclusions
Risk of future violence is a very important forensic task. However, risk assessment is an emerging field. Many tools are in the developmental or research stages. It is likely that several tools will be used simultaneously because each uses a slightly different risk perspective and may provide unique information. Comparing tools allows a practitioner to choose the tool that best suits his/her population and situation. Once dangerousness is estimated, there are treatments, which if applied with sufficient intensity and length of time, can be effective in reducing future risk of violence. It can also help with lesser security and release determinations. We have the tools and they are widely used in Canada. They are finding more acceptance in the US, as well.
References on request
Dr. Kathryn Seifert is a psychotherapist with over 30 years experience in mental health, addictions, and criminal justice work. Dr. Seifert has authored the CARE and numerous articles. She speaks nationally on mental health related topics and youth violence. She is an expert witness in the areas of youth and adult violence and sexual offending. Her latest book is coming soon: How Children Become Violent. For more information go to http://www.drkathyseifert.com