Reading time: 3 minutes, 17 seconds
Over the years, violence risk assessment has evolved from descriptive, unstructured clinical judgment to predictive, actuarial risk assessment formulas and now to explanatory, structured professional judgment that helps clinicians formulate risk management safety plans.
During the International Forensic Psychiatry Lecture Series, Dr. Michael Doyle discussed the three generations of violence risk assessments and his recommendations for forensic mental health professionals. (You can watch a recording of his March 27, 2025, lecture online.)
First Generation: Unstructured Clinical Judgment
The first violence risk assessments were unstructured, discretionary, and descriptive. The clinical judgment was subjective, relying on the personal and professional experiences, biases, and intuition of the mental health professionals.
The results were only slightly better than chance.
These assessments were often overly influenced by political and institutional caution, as little benefit was seen in releasing somebody early.
Contrary to public perception, people with mental illness account for a very small percentage of violence in society, and then it may not be due to the illness. That said, people with mental illness have a higher risk for violence when they:
- have an underlying personality disorder or trait;
- have early trauma and victimization, or an upbringing that condones violence;
- misuse substances, particularly alcohol and stimulants; and
- have paranoid symptoms or dysregulated anger.
Another problem with the first generation of violence risk assessments was that dangerousness was not well-defined. Although in 1975, the Butler Committee on Mentally Disordered Offenders defined dangerousness as “an individual’s propensity to cause serious physical injury or lasting psychological harm.”
Second Generation: Actuarial Risk Predictions
Research found that predictions of violence could be improved by using validated statistical predictors of violence. This led to the second generation of risk assessments, which incorporated data on a wide range of actuarial risk markers to calculate the risk.
This actuarial approach of predicting violence was usually not discretionary but an empirically based, structured, and objective way to calculate risk. Examples include the Classification of Violence Risk (COVR), Psychopathy Checklist: Screening Version (PCL:SV), and Violence Risk Appraisal Guide (VRAG).
“If you look at the … clinical judgments versus the actuarial predictions, the actuarial predictions would almost always be much more accurate in large groups in determining which people in which groups would be violent in the future,” Dr. Doyle said. However, this comparison does not account for the risk paradox in clinical practice, where assessors must take action to minimize the risk of violence when they predict violence will occur, thereby proving their prediction wrong.
Third Generation: Structured Professional Judgment
Dr. Doyle highlighted that violence risk assessments need to be both evidence based and reflect professional and legal considerations. They not only needed to be structured but also flexible and responsive to the uniqueness of each individual being assessed.
The challenge becomes combining the actuarial approach with professional clinical judgments. This has been described as a third generation of risk assessment, facilitated by structured professional guidelines.
These structured professional judgment guidelines move the emphasis from prediction to prevention and risk management. They are based on sound scientific evidence yet are practical and relevant to clinical practice. They are used as tools to support the mental health professional’s decision, to help prevent — not predict — violence and guide risk management strategies.
Examples include the Historical, Clinical, Risk Management (HCR-20) and Structured Assessment of Protective Factors for Violence (SAPROF).
Recommendations
Dr. Doyle recommends forensic mental health professionals take an assets-based approach to managing risk. This means asking what makes people safe, rather than just what makes them a threat. This shift moves risk assessments from identifying risk factors and predicting outcomes to managing and preventing violence, focusing on rehabilitation and enhancement of internal, motivational, and external strengths.
The goal is to provide effective care and treatment to reduce the risk, promote recovery, and support patients moving to lower levels of security or to successfully re-establish themselves in the community with suitable safeguards.
This means moving from the defensive approach to defensible decision making, which can require risky solutions.
Defensive actions are overly cautious, with clinicians prioritizing self-protection from potential legal or professional repercussions over focusing on their patients’ best interests.
Defensible actions and decisions can be justified using evidence-based practices, established professional standards of care, ethical principles, and the patient’s needs.
This change can mean risky solutions, such as:
- Taking positive risks crucial for recovery, rehabilitation, and autonomy
- Encouraging patients to do activities that promote their well-being, even if those activities carry some inherent risks
- Encouraging a patient to live independently or return to work, even if it carries some risk, because it can enhance their mental and emotional well-being
- Fostering care plans that reflect personal goals and preferences
“Therapeutic risk taking is not therapeutic if the risk is not understood nor if the action planned is not defensible,” Dr. Doyle said in closing.
Written by Virginia St-Denis and approved by Dr. Michael Doyle.
