Download Clinical Risk Assessment - Dr Brink

Clinical Risk Assessment
Wallace Brink
StR Forensic Psychiatry
Langdon Hospital
Student who
killed mother
and unborn
twins sent to
Rampton
Saturday Telegraph
May 6th 2006
Definitions





Risk: the likelihood of an adverse event.
Risk Factors: features associated with increased
risk.
Risk Assessment: an estimation of the
likelihood of particular adverse events occurring
under particular circumstances. Within a
specified period of time.
Risk Formulation: organisation of the risk data
to facilitate risk management.
Risk Management: organised attempts to
minimise the likelihood of adverse events
Risk




Depends on the individual and the context
Objective
Dynamic
Not equal to DANGEROUSNESS
Not dangerous 1---5---10 Very dangerous
Types of risk assessment
 Clinical
assessment
 Unstructured or clinical
 Structured (e.g. HCR 20)
 Actuarial approach
Clinical Risk Assessment
Awareness that risk is dynamic
 Adopt a structured approach
 Explicit working
 Consider protective factors as well as risk
factors

Clinical Risk Assessment
Gather necessary information
 Keep good records
 Communicate your assessment
 Base your interventions on the risk
assessment

Practical and Systematic

Gather information from:
 The
individual being assessed
 Others who know them
 Records
 Take a full history
Consider the risks involved









Is there a risk of harm?
What sort of harm?
What degree?
Who is at risk?
How likely is it that harm will occur?
What is its immediacy?
How long will the risk last?
What are the factors which contribute to the risk?
How can the factors be modified or managed?
What is the relationship between
risks?
Absconding
Non compliance
Substance use
Mental state deterioration
Physical assault
McNeil et al 2003

Clinical factors may be most relevant for the
estimation of short term risk in acutely ill
patients

Historical factors may be most relevant for
estimating the long-term risk in treated patients
ECA Study: Swanson 1990

Major mental disorder: 5 fold increase in
violence compared to those without major
mental disorder (10-13% verses 2%)

Substance misuse: 10 fold increase in violence
compared to non-drug users (19-35% verses
2%)
Birth Cohort Study: Hodgins (1992)

Odds Ratio of 4 for violence among men with
major mental illness compared with controls

Odds Ratio of 27 for violence among women
with major mental illness compared with
controls
Other factors associated with
violence

Male gender, young age, low socio-economic
status


Male gender, young age, low educational level


Swanson, 1990
Link, 1992
Discharge to poverty

Silver et al 1999
MacArthur Violence Risk
Assessment Study: Steadman 1998

Prospective 1 year follow up of 1000 discharged
patients compared to community controls for
levels of violence

No association found between mental illness
and violence

May be indication of the success of risk
management
Summary of violence literature



Substance misuse is a major risk factor with or
without mental disorder
Socio-demographic factors contribute
significantly
Contribution of mental illness is relatively small
Accuracy of clinical assessment

Link 1993: predictions in emergency room
patients
correct 1 in 2 attempts
 clinicians significantly underestimated risk in women
 if used just the historical data on the same patients
the sensitivity increased at the expense of the
specificity

Mulvey and Lidz 1998



Asked doctors to predict which of the patients
assessed in the ER would be violent
The clinicians did reasonably well in predicting
place, target, severity of violence and
involvement of alcohol in violence
Clinicians overestimated the influence of noncompliance and drug misuse upon risk of
violence
Violence is relatively rare and
consequently accurate prediction is
difficult
Monahan grid
Personal / dispositional (static)
 Historical (static)
 Contextual (dynamic)
 Clinical (dynamic)

Personal / dispositional
Demographic
 Personality
 Neuropsychological
 Physical

Historical
Family and personal history
 Work and education
 Psychosexual development
 PPH and PMH
 Previous offending and antisocial
behaviour

Contextual
Level of support and supervision (actual
and perceived)
 Availability of victim / weapons /
substances
 Perceived stress
 Interests (sexual, violence, cruelty, racial)

Clinical
Delusions, hallucinations, passivity
 Depression, mania
 Anger/rage, impulse control
 Paranoid disposition, jealousy
 Fantasies
 Personality disorder
 Substance use

Risk of violence to others:
victim
Relationship to perpetrator
 Particular characteristics
 Vulnerability
 Availability

Mullen’s approach
Mullen P. Dangerousness, Risk and the
Prediction of Probability. The New Oxford
Textbook of Psychiatry. (Eds M.G. Gelder, J.J.
Lopez-Ibor and N.C. Andreasen). Chapter
11.4.3. Oxford.
Pre-existing vulnerabilities
Increase
 Male
 Young
 Disrupted or abusive
Childhood
 Antisocial
 Suspicious
 Impulsive
Irritable
Decrease
 Over 35 years of age
 Good pre-morbid
personality
 Stable/nurturing
childhood
 Sensible
Social and Interpersonal factors
Increase
 Poor social network
 Lack of education
 Lack of work skills
 Rootless
 Poverty
 Homelessness
Decrease
 Good social network
 Stable accommodation
 Employment
 A confidante
 Supportive intimate
relationship
Mental Disorder
Increase
 Active symptoms
 Poor compliance
 Poor engagement with
services
 Treatment resistance
 Lack of insight
Decrease
 Absence of active
symptoms
 Good compliance
 Good engagement
 Good treatment
response
 Good insight
Substance Misuse
Increase
 Present
Decrease
 Absent
State of Mind
Increase
 Anger/fear
 Threats
 Delusions








Evoking fear
Provoking indignation
Provoking jealousy
Involving jealousy
Involving injury/threat from close
relative or companion
Clouding consciousness and
confusion
Ideas of influence
Command hallucinations
Decrease
 Amotivational
Situational Triggers







Availability of weapons
Loss
Demands and expectations
Confrontation
Change
Physical illness
Other provocation
Good risk assessment








Reviewed on a regular basis
Reviewed if there are new concerns
Multi-disciplinary
In collaboration with the patient and their carer
Limitations of your assessment noted
Includes factors which reduce risk of future violence
Only useful if disseminated
Informs the management plan
Define the risk


Severity best predicted by prior violence
Imminence best predicted by
pattern of violence
 statements,
 life circumstances.


Likelihood best predicted by actuarial models
Dvoskin and Heilbrun 2001
Homicide Inquiries: why do things
go wrong





Failure to lend sufficient weight to reports by carers and
members of the public about disturbed behaviour
An undue emphasis on the civil liberties of patients at
the expense of increased risk of suicide or of violent
behaviour
A failure to properly implement the MHA
A tendency to take cross-sectional rather than longterm view of the risk of suicide or violence
A failure to share information in the best interests of
the patient
W v Egdell [1990]

Duty of confidence to the patient is not
absolute

Balance between the interest in confidentiality
and in public safety
Thoroughness






Attention to detail
Accurate and detailed record keeping
Comprehensive history taking
Avoid minimising incidents
Linking incidents
“Asking the unaskable”
Multi Agency public protection
arrangements MAPPA

Offenders who pose a risk of serious harm to
others
Level 1: Caused serious harm previously, manageable
by a single agency
 Level 2: Pose a serious risk to others but not an
imminent risk
 Level 3: Pose and imminent and serious risk

Interagency working






Healthcare
Social Services
Housing departments
Police
Probation
Day centres/hostels
The defendable decision







Take all reasonable steps
Use reliable assessment methods
Seek information you do not have
Thoroughly evaluate all relevant information
Stay within agency policies and procedures
Record and account for decision making
Communicate the plan to others involved
Risk Management: CPA




Actions to minimise the hazards
Actions to enhance protective factors
Review date
Contingency plan to include



Arrangements for when the co-coordinator is unavailable
Arrangements for when part of the care plan can not be
provided
Crisis plan to include:

Action to be taken if mental state is rapidly deteriorating
Positive risk management involves



Weighing up the potential benefits and harms
Plans which support the positive potentials and
minimise the risks
An element of risk because the potential
positive benefits outweigh the risks
HCR-20
ASSESSING RISK
FOR VIOLENCE
VERSION 2 - 1997
Christopher D. Webster
Kevin S. Douglas
Derek Eaves
Stephen D. Hart
Mental Health, Law, and Policy Institute
Simon Fraser University
Scope & Purpose
“The main aim was to produce a guide which
would be rooted in scientific knowledge… be
defined precisely,… and be designed for
efficiency with time constraints in mind”
General Principles for Improving
Prediction Accuracy





“What exactly is the referral question?”
“Opinions formed about risk under one set of circumstances
(e.g., risk for violence in the community) may have limited
pertinence to another set (e.g., violence while
institutionalised).”
“Clinicians who have been seeing patients for psychotherapy
may wish to decline offering assessments of risk for such
patients”
“Very hurried or pressured assessment, or those based on
partial information, invite inaccuracy”
“The scientific knowledge from which the assessment is
formulated should be current”
General Principles for Improving
Prediction Accuracy 2



“The particular scheme chosen should correspond as closely as
possible to the type of population from which the assessee is
drawn”
“Whenever possible, the base-rate of violence in pertinent
populations should be obtained or estimated. It is important
that this base rate, which may be quite low in some
populations, guide the eventual statement of risk”
“Particular importance should be ascribed to historical
considerations, which should anchor such modifications as
might be suggested by analyses of clinical and situational
factors… Cross-checking of information is crucial at every
step”
Organisation of the HCR-20

“An important aspect of the HCR-20 is that it includes
variables which capture relevant past, present, and future
considerations. Historical, or static factors are weighted as
heavily as the combined present clinical and future risk
management variables”

20 item structure:



Historical – 10 past history factors
Clinical – 5 present variables
Risk Management – 5 future issues
Administration

“Current research is revealing the necessity of multiple sources
of information in making risk assessments… A thorough and
thoughtful review of all available files must be completed”

“Assessors ought to include in their reports all sources which
they did consult, did not consult, or were unable to consult”
Defining Violence

“Violence is actual, attempted, or threatened harm to a person
or persons… Violence is behaviour which obviously is likely
to cause harm to another person or persons… In a general
sense, then acts which are serious enough to result in criminal
or civil sanctions, or for which the perpetrator could have been
charged, should be considered violent, and those that are not as
serious as this should not be considered violent… All sexual
assaults should be considered violent behaviour.”
Historical
(Past)
Clinical
(Present)
H1. Previous Violence
H2. Young Age at First
Violent Incident
H3. Relationship
Instability
H4. Employment
Problems
H5. Substance Use
Problems
H6. Major Mental Illness
H7. Psychopathy
H8. Early Malajustment
H9. Personality Disorder
H10. Prior Supervision
Failure
C1. Lack of Insight
C2. Negative Attitudes
C3. Active Symptoms of
Major Mental Illness
C4. Impulsivity
C5. Unresponsive to
Treatment
Risk Management
(Future)
R1. Plans Lack
Feasibility
R2. Exposure to
Destabilisers
R3. Lack of Personal
Support
R4. Noncompliance
with Remediation
Attempts
R5. Stress
Coding Items
No
The item definitely is absent or does not apply.
Maybe The item possibly is present, or is present only to a limited
extent.
Yes
The item definitely is present.
Omit Don’t Know – There is insufficient valid information to
permit a decision concerning the presence of absence of
the item.
H1. Previous Violence
+/-
+
No previous violence
Possible / less serious previous violence (one or two acts of moderately
severe violence)
Definite / serious previous violence (three or more acts of violence, or
any acts of severe violence)
The scoring scheme here is intended to capture the density of
previous violence. For this reason the number of past violent acts
is combined with the severity of past violence to determine the
score… All violence which occurs up to and including the time of
assault is included as “previous violence”.
H2. Young Age at First Violent Incident
+/+
40 years and older at first know violent act
Between 20 and 39 years at first know violent act
Under 20 years at first known violent act
We are aware that, in general, the younger a person was at his or
her first act of violence, the greater is the probability of future
violence… Age is established by considering the date of the first
known violent incident, and not using the date of the index offence
or assessment.
H3. Relationship Instability
+/+
Relatively stable and conflict-free relationship pattern
Possible / less serious unstable and / or conflictual relationship pattern
Definite / serious unstable and / or conflictual relationship pattern
This item applies only to ‘romantic’, intimate, or non-platonic
partnerships, and excludes relationships with friends and family.
The item is geared toward whether an individual show evidence of
having the ability to form and maintain stable long-term
relationships, and engages in these when given the opportunity.
“Instability” may show in several ways: many short-term
relationships; absence of any relationships; presence of conflict
within long-term relationships.
H4. Employment Problems
+/+
No employment problems
Possible / less serious employment problems
Definite / serious employment problems
Individuals who warrant a high score on this item may refuse to
seek legitimate employment, or have a history of having many
jobs within short-term periods, or of frequently being fired or
quitting. The primary focus of this item is the presence or absence
of employment problems.
H5. Substance Use Problems
+/+
No substance use problems
Possible / less serious substance use problems
Definite / serious substance use problems
The assessor is interested in whether there exists impairment of
functioning in areas of health, employment, recreation, and
interpersonal relationships which is attributable to substances.
H6. Major Mental Illness
+/+
No major mental illness
Possible / less serious major mental illness
Definite / serious major mental illness
A diagnosis of major mental illness should conform to an official
nosological system such as the DSM-IV or ICD-10. This item is
scored on the basis of past history and is unaffected by whether the
disorder is currently active or in remission. This item applies to
illnesses involving disturbances of thought and affect (e.g.,
psychotic illnesses, manic mood illnesses, organic illnesses,
retardation, etc.).
H7. Psychopathy
+/+
Nonpsychopathic
Possible / less serious psychopathy
Definite / serious psychopathy
It must be stressed that this rating is to be made on the basis of an
informed and trained psychopathy assessment using the PCL-R or
PCL:SV.
It may be appropriate to modify the scoring ranges according to
local (e.g. UK) populations.
H8. Early Maladjustment
+/+
No maladjustment
Possible / less serious maladjustment
Definite / serious maladjustment
This item includes two very different ways in which childhood
maladjustment predicts later violence. One way is through
childhood victimisation, the other through being a childhood
victimiser… Although both factors predict adult violence, they
clearly have different implications for intervention.
H9. Personality Disorder
+/+
No personality disorder
Possible / less serious personality disorder
Definite / serious personality disorder
A diagnosis of personality disorder should conform to an official
nosological system such as the DSM-IV (APA, 1994), or the ICD10 (WHO, 1992).
H10. Prior Supervision Failure
+/+
No supervision failure(s)
Possible / less serious supervision failure(s)
Definite / serious supervision failure(s)
Failures during any institutional or community placement are
relevant here. A supervision failure is considered to be serious if it
resulted in the individual being (re-)apprehended or (re-)
institutionalised by a correctional or mental health agency.
Clinical Items

“Although historical items have the strongest
support in terms of predictive acumen, there is
no dearth of well-established clinical
constructs that may be relevant to the
assessment of risk.”
C1. Lack of Insight
+/+
No lack of insight
Possible / less serious lack of insight
Definite / serious lack of insight
This item refers to the degree to which the assess fails to
acknowledge and comprehend his or her mental disorder, and its
effect on others.
C2. Negative Attitudes
+/+
No negative attitudes
Possible / less serious negative attitudes
Definite / serious negative attitudes
We here refer to the kind of pro-criminal and antisocial attitudes
that have some likelihood of eventuating in violence.
C3. Active Symptoms of Major
Mental Illness
+/+
No active symptoms of major mental illness
Possible / less serious active symptoms of major mental illness
Definite / serious active symptoms of major mental illness
Assessors should follow a classification system, such as the DSMIV (APA, 1994) or ICD-10 (WHO, 1992).
C4. Impulsivity
+/+
No impulsivity
Possible / less serious impulsivity
Definite / serious impulsivity
Impulsivity refers to dramatic hour-to-hour, day-to-day, or weekto-week fluctuations in mood or general demeanour… Impulsive
persons are quick to (over-) react to real and imagined slights,
insults, and disappointments.
C5. Unresponsive to Treatment
0
1
2
Responsive to treatment
Possible / less serious unresponsiveness to treatment
Definite / serious unresponsiveness to treatment
This item includes any treatment designed to ameliorate criminal,
psychiatric, psychological, social, or vocational problems. It does
not refer to treatments which are largely irrelevant to criminal or
psychiatric tendencies.
Risk Management Items

“This section centres on forecasting how
individuals will adjust to future circumstances.
Although admittedly speculative, the exercise
serves to stimulate development of appropriate
risk management plans.”
R1. Plans Lack Feasibility
+/+
Low probability that plans will not work
Moderate probability that plans will not work
High probability that plans will not work
Lack of feasibility may be due to the fact that community agencies
are unwilling or unable to provide assistance. Alternatively, the
patient may have played no role in making plans or be uninvolved
with peers or family.
R2. Exposure to Destabilisers
+/+
Low probability of exposure to destabilisers
Moderate probability of exposure to destabilisers
High probability of exposure to destabilisers
In large part, persons may be exposed to destabilisers because of
inadequate professional supervision.
R3. Lack of Personal Support
+/+
Low probability of lack of personal support
Moderate probability of lack of personal support
High probability of lack of personal support
This item can be coded present if support (emotional, financial, or
physical) from friends or family is unavailable, or if such support
is available but the individual is unwilling to accept it.
R4. Non-compliance with
Remediation Attempts
+/+
Low probability of non-compliance with remediation attempts
Moderate probability of non-compliance with remediation attempts
High probability of non-compliance with remediation attempts
Individuals who score high on this item may lack motivation to
succeed and willingness to comply with medication and therapy,
or refuse to follow rules.
R5. Stress
+/+
Low probability of stress
Moderate probability of stress
High probability of stress
This item can be coded present if the individual is likely to be
exposed to serious stressors. Alternatively, the anticipated
stressors may be less serious, but the assessor is concerned that the
individual will cope poorly with them.
HCR 20: scenarios








Nature
Motivation
Victims
Severity
Imminence
Frequency
Duration of risk
Likelihood






Risk-enhancing factors
Risk-protective factors
Monitoring
Treatment
Supervision
Victim safety planning
References









Guideline for clinical risk assessment and management in
mental health services. Ministry of Health (New Zealand) 1998
Dangerousness, Risk and the Prediction of Probability. Mullen P.
The New Oxford Textbook of Psychiatry. (Eds M.G. Gelder, J.J. Lopez-Ibor
and N.C. Andreasen). Chapter 11.4.3. Oxford.
The state of contemporary risk assessment research. Norko MA
and Baranoski MV. Can J Psychiatry (50) 1, 18-26.
Best Practice in Managing Risk. Department of Health June 2007
Rethinking risk to others in mental health services. Final report
of a scoping group. June 2008. RCPsych.
‘Giving up the Culture of Blame’Risk assessment and risk
management in psychiatric practice. February 2007. RCPsych
Risk assessment. A word to the wise. Vinstock M. APT (1996) 2, 310
Evaluating risks. Kapur N. APT (2000) 6, 399-406
Assessing risk of interpersonal violence in the mentally ill.
Mullen P. APT (1997) 3, 166-173.
“Prediction is very difficult,
especially about the future”
Niels Bohr (1885-1962)