Dr adrian keller psychosis and violence vietnam conference presentation final version

VIOLENCE AND MENTAL ILLNESS – RISK ASSESSMENT AND MANAGEMENT Dr Adrian Keller Forensic Psychiatrist Justice Health, NSW SUMMARY The relationship between psychosis and violence The role of violence risk assessment in persons with psychosis Managing violence risk in persons with psychosis “Most people who are violent are not mentally ill…and most people who are mentally ill are not violent.” Cycle of Violence Psychosis and violence – the relationship  Relative risk of violence is increased in persons with psychosis - Odds ratio of – (males) - – 29 (females) o Comorbid substance abuse substantially increases the risk (4 x greater than for persons with psychosis and no substance abuse) o However, the risk of violence in persons with both psychosis AND substance abuse is no higher than for substance abuse alone o Homicide rates for persons with psychosis are strongly correlated with the total homicide rates within the general community of that country – a range of non-specific risk factors, which vary in proportion from country to country, mediate the risk of homicide amongst persons with psychosis These include: - Substance abuse - History of childhood conduct disorder - Lower socioeconomic status - Being a victim of violence ‘Negative symptoms’ (e.g amotivation and blunted affect) appear to be a protective factor against violence in persons with psychosis Psychotic Symptoms and Violence: Recent Findings Hallucinations and violence: 17 studies reviewed  No evidence that auditory command hallucinations increase violence risk  Several studies found a link between command hallucinations that contain violent content, and violence  No evidence that co-occurrence of hallucinations and delusions increases risk for violence Delusions and violence: 20 studies reviewed  Most studies find a link between delusions and violence  13 out of 14 studies found that persecutory delusions increase risk for violence  of studies found that threat/control-override symptoms (TCO) are associated with violence  out of studies found that “delusional distress” ( delusions accompanied by anger, fear/anxiety or suspiciousness/hostility) was associated with increased risk for violence o Relative risk of homicide in person with psychosis is high (15 – 20 x general population), although ‘absolute risk’ is low (0.3% vs 0.02% in general population) o Rates of homicide have not changed appreciably before and after ‘deinstitutionalisation’ o Shifting focus from the ‘relative’ to the ‘absolute’ risk of violence posed to the community may reduce the stigma associated with a diagnosis of psychosis Homicide and other forms of violence are far more common in persons with psychosis who are not receiving adequate treatment Those patients who have “never received treatment” appear to be at higher risk than patients who “have a record of poor adherence to medication despite previous treatment” The DUP (duration of untreated psychosis) also seems to be correlated with violence Victims of homicide committed by persons with psychosis Random attacks on strangers is uncommon Most common victims are close family (e.g parents, spouse, child) Next most common are close associates and friends For those patients in hospital or custody, proximity of victim is important factor (e.g mental health professionals; other patients or other prisoners) Understanding Risk  RISK: The probability of an outcome  RISK FACTOR: A factor that has an association with a risk  RISK ASSESSMENT: The process of evaluating the risk  RISK PREDICTION: A statement of the probability of an outcome  RISK MANAGEMENT: The process through which risk is contained The Nature of Risk Important principles to consider in relation to the nature of risk:  Risk changes continually A person’s risk can change rapidly and over the course of a short period of time, as well as over a longer period of time  Risk cannot be eliminated Risk can be managed when mental health professionals engage in ‘responsible risk taking’  Risk management begins with assessment and identification of empirically known and idiosyncratic risk factors – and ends with a plan to manage those risk factors P r o te c tiv e F a c to r s I R is k F a c to r s Risk Over Time D y n a m ic T im e > B a s e li n e Factors Correlated With Risk for Violence in the Mentally Ill (MacArther 2000) GENDER  Mentally ill men are at no more risk of violence than mentally ill women  Male violence is typically more serious than female violence  Women more often violent in the home  Men more likely to have associated alcohol or illicit substance misuse APPROACHES TO RISK ASSESSMENT Why Have an Approach  Adopting a consistent approach enhances reliability  An empirically based, documented risk assessment may reduce the probability of a hazard occurring in an individual patient  Clarifying the parameters of the risk and management plan establishes an understandable rationale for the decision taken, and provides the assessor with a defensible position when a hazard does occur (as, over time, it inevitably will) Approaches to Risk Assessment The Extremes – Flexible & Rigid  Unstructured Clinical Judgment (no factors and no structure)  Actuarial Approach(specific factors and rigid structure) Unstructured Clinical Judgment ADVANTAGES DISADVANTAGES  Evaluator determines  Naïve , impressionistic, subjective, intuitive  Evaluator exercises  The “in my experience” approach  Akin to making a diagnosis without knowledge of the signs and symptoms that are associated with the condition how and what information is gathered decision making discretion  Most common method used by mental health professionals  Minimal cost  Minimal time required Actuarial Approach DISADVANTAGES ADVANTAGES   May not be generalisable to your population  Factors are mostly static and not account for other risk reducing factors and circumstances  Based on outcomes in context of strict research conditions  Can identify optimal cut off points to improve predictive accuracy Tells you the risk group and less about the individual  Transparent and defensible Becomes the focus at expense of clinical experience  Does not assist in management  Too specific: X% in B population over T time Use of an equation, graph, table to provide a probability estimate of risk for a particular risk group  Better inter-rater reliability  Have scientific validity and statistical accuracy    Easy to use, less skill, checkbox Structured Professional Guidelines ADVANTAGES DISADVANTAGES  Utilizes a list of empirically based risk factors vetted by a experts   Operationalises these risk factors  Provides a fixed scoring guide   Provides guidance for decision making  Assists in risk categorization  Makes the assumption that ultimately clinical discretion needs to be applied in individual cases  Easy to use and can be applied by most clinicians  More generalizable than actuarial methods    Does not take into account factors specific to the individual case Some focus on risks and weaknesses, not strengths Focuses on only one type of risk – violence Attempt to predict risk in longer term rather than shorter term Focus on forensic populations HCR 20 Historical Factors  Static  Tell you the base level of risk  The best you can Clinical  Current functioning  Dynamic factors  What needs to be done now Risk Factors  Future Issues to consider  What need to be done in the future The HCR-20: a Structured Clinical Rating Scale for Risk for Future Violence (Webster et al., 1997) Historical items: Clinical variables: (1) Previous violence (1) Lack of insight (2) Young age at first violent incident (2) Negative attitudes (3) Relationship instability (4) Employment problems (3) Active symptoms of major mental illness (4) Impulsivity (5) Unresponsive to treatment (5) Substance use problems Risk management items: (6) Major mental illness (1) Plans lack feasibility (7) Psychopathy (2) Exposure to destabilizers (8) Early maladjustment (3) Lack of personal support (9) Personality disorder (4) Non-compliance with remediation attempts (10) Prior supervision failure (5) Stress THE “HOLY GRAIL” (a) identification of causal, dynamic/ variable risk factors (b) that are amenable to treatment The KEY QUESTION remains: Having regard to the risk factors relevant in this case, can this person be safely managed in the expected environment? (e.g community) Remember: Low Risk DOES NOT = No Risk TAKE HOME MESSAGES  The presence of psychosis increases a person’s risk of violence – but remember: whilst relative risk is high, absolute risk is low  Violence is a complex, multi-factorial phenomenon and most risk factors are common to persons with/without psychosis The most important of these is substance misuse  The highest risk of violence associated with psychosis may be in the first episode of illness, prior to the commencement of treatment  The most rigorous and clinically appropriate approach to violence risk management is the use of structured professional guidelines ... between psychosis and violence The role of violence risk assessment in persons with psychosis Managing violence risk in persons with psychosis “Most people who are violent are not mentally ill and. .. factor against violence in persons with psychosis Psychotic Symptoms and Violence: Recent Findings Hallucinations and violence: 17 studies reviewed  No evidence that auditory command hallucinations... are mentally ill are not violent.” Cycle of Violence Psychosis and violence – the relationship  Relative risk of violence is increased in persons with psychosis - Odds ratio of – (males) - – 29