Could we have done more? Risk assessment and violence

Every week, we read of a violent act that could have or should have – or we like to think that it could have or should have – been prevented. The perpetrator is usually a repeat offender. He (for usually it is a male) has been granted bail, granted parole, given a day pass or moved to a less secure facility, discharged to a half-way house or simply to the community, from either a correctional system facility or a forensic psychiatry facility.

In most cases, a review or parole board has sanctioned the change to a less secure environment with less supervision. Many of these decisions, in part, have been based on a “risk assessment” by mental health professionals (forensic psychiatrists or psychologists), and by an assessment of the behaviour of the individual while in custody or treatment. This latter often includes comments on “admitting guilt and taking responsibility,” “showing remorse,” “engaging in pro-social activity” and “attending therapy or counselling.”

Currently, some objective tools (evidence-based psychopathy checklists) are used in a risk assessment but some subjective judgements are based on self-reporting, interviews and testimony. Ultimately though, a risk assessment will arrive at very broad conclusions such as “low risk to re-offend,” “moderate risk” and “high risk.”

In any decision to discharge, release or reduce security, legal, humanitarian and civil liberty considerations must play a role.

But there is a different and far more objective question that could be asked of the mental health professionals: “Apart from incarceration, with this individual, do we have the tools to reduce to zero or minimal risk the possibility of re-offending?” And this question can be objectively answered based almost solely on the original crime(s) or “index offences,” the context of those crimes and the presence or absence of a treatable condition.

It is a question of specificity applied to five elements of the index crime or crimes. These elements are:

• Victim(s). Was the victim (or victims) a specific person in the life of the perpetrator, e.g., mother or wife (very specific) vs. young woman in a park?
• Context of index offence. Did the violence occur within a specific context or a random encounter, e.g., was the perpetrator in a multi-year relationship with the victim or had they recently met in a bar, during a drugs for money exchange or passing in the street?
• Illness or condition. Does a specific illness or identifiable condition account for the act of violence, e.g., schizophrenic delusion or manic episode vs. “psychopathic traits?”
• Treatment. Do we have a specific (and effective) treatment for this illness or condition, e.g., anti-psychotic medication vs. anger management counselling?
• Compliance. Do we have a specific way of monitoring the illness and compliance with treatment, e.g., physiological test or injectable medication vs. self-reporting?

The answers to these questions can determine if we actually have the tools that could substantially reduce risk. The risk can be quantified with scores of 1 to 5 applied to each element, with 1 being very specific and 5 being non-specific:

Victim: Mother would score 1 as very specific; girl in the park would score 5.

Context: Multiyear dependent relationship would score 1; person met in a bar 5.

Illness: Schizophrenia would score 1 or 2; psychopathic tendencies 4 or 5.

Treatment: Antipsychotic medication would score 1 or 2; counselling 5.

Compliance: Monthly injectable medication would score 1; self-reporting 5.

This method avoids several of the problems and mistakes that can be found in the current decision-making processes. These mistakes include over-reliance on self-reporting and subjective observation; mental health workers and others developing (quite naturally and predictably) empathy for the perpetrator; relying on non-specific “treatments” that are not actually effective; and relying on observations of human behaviours that are easily faked or fleeting (self-reporting, remorse, acceptance of responsibility, finding religion, good attendance, pro-social attitudes, etc.)

The following case examples illustrate the problem of relying on the current process:

1. In Hamilton, a man in his 40s, suffering from schizophrenia, living in a boarding home, stops taking his medication, becomes psychotic and kills the female supervisor/landlady of the boarding home. It turns out that he had killed his mother in his early 20s, had been found not criminally responsible and then treated for some years within forensic psychiatry services before being discharged to supported living in the community.

The system had inadvertently recreated the specific conditions of the index offence: multi-year dependent relationship with a mother-figure; non-compliance with a specific and effective treatment.

2. A man is convicted of rape and murder of a young woman, a stranger to him. He is found not criminally responsible and goes into the forensic psychiatry system. Looking through the old records, one can find that he is suspected of two other similar offences but not prosecuted for these. His diagnosis is “psychopathic traits.” On the basis of good participation in programs, group sessions, anger management sessions, displays of remorse and acceptance of responsibility while first in a maximum-security facility and then a medium-security facility, he is put forward for transfer to a minimum-security facility and release. He is rejected by the minimum-security facility and shortly after, on an unsupervised day pass, rapes a girl in the nearby park and throws her off a small bridge.

Applying the system outlined above would tell us that this man scores 23 to 25 (very high risk), and that we do not actually have specific tools, short of incarceration and/or 24/7 supervision, that would reduce this risk.

3. Vince Li, suffering from schizophrenia and delusional, kills a passenger on a Greyhound bus. He is found not criminally responsible, incarcerated and treated successfully with anti-psychotic medication. He is given increasing freedom and then eventually discharged to the community with no restrictions. His final discharge is based, in part, on the success of treatment and his exemplary behaviour, including insight and remorse.

In this example, the victim and context were not specific, but the illness and its treatment as well as the monitoring compliance with the treatment are very specific. And, predictably, Li presents minimal risk of re-offending while maintained on anti-psychotic medication.

But Li’s final discharge fails to take into account what we know of human behaviour and of schizophrenia and delusions. That is, many, if not most patients, at some point, go off their medications no matter their intentions and awareness. Many different unpredictable circumstances may cause this. Many, if not most patients who stop their medication, do not tell their doctors or nurses. When a person with schizophrenia relapses, his or her symptoms return in the same form, even years later. That is, upon relapse, the same or a very similar delusion returns.

Thus, Vince Li continues to pose some risk. If he stops his medication, his illness will relapse and the same delusion that previously led him to kill a stranger on a bus will return. The victim and context in this case are easily replicated. This risk could have been eliminated by prescribing and enforcing and supervising compliance on medication for life. Such life-long monitoring of compliance with medication does not, to me, seem excessively onerous considering the index offence and the potential severity of a repeat offence if relapse occurs.