RISK ASSESSMENT
LECTURE 1
1. Case study – the corn-killing
court psychiatrist assessment:
- introverted (not part of the central 8/ non-criminogenic need)
- tendency to bottle up aggression (not part of the central 8/ non-criminogenic need)
- difficulty empathizing with the feelings of others (part of the central 8)
- no personality disorder (it was not mentioned how the psychiatrist came to that
conclusion/ not part of the central 8)
- anger issues regarding his mother (not part of the central 8)
- very high chances for recidivism unless treatment is provided
! the risk of recidivism should be decided first, and then the need for an intervention
should be established
- no guilt/ no remorse (not part of the central 8)
2. History of risk assessment
older and newer definitions
- they portray different reasons for risk assessment
- the first one points to the severity of the crime – it is not that important
- the second definition – risk assessment is not only determining risk but also identifies the
need of interventions
several definitions are available
risk management – the chance of re-offending is never 0, but what can be done to reduce
the risk factors a person has?
why do we do risk assessment?
- increase safety
- public interest
- ethical importance
- therapeutic importance
- communication
Lombroso (not completely wrong)
- criminal behavior is inborn and it is shown in one’s physical characteristics
- characteristics related to higher chances of becoming violent
father of risk assessment – Burgess
- created the first risk assessment tool
- the tool was actuarial in nature
- re-offending rates dropped
- data from offenders who were on parole
- the instrument has been validated
Baxstrom v Herold (1966) – American case
- Baxstrom was released and many high-risk patients were let back into society or sent to
regular mental institutions (not forensic psychiatric hospitals)
expertise of the expert study
- teachers who had no background in forensic psychiatry/ law
- low IRR between the teachers and the specialists
- many similarities but the teachers did better than the psychiatrists
, - the more information that is available, the more conservative one is in their judgment
- psychiatrists guess correctly 1/3
the 90s
- the Dutch TBS system was praised across the world
- denial/ lack of empathy – risk factors
- the experts were wrong when pointing out important risk factors
Monahan – investigated true risk factors and the RNR model was developed later
2.1 RNR model
4 main principles (22 in total):
1) evidence-based practice
2) risk principle – who to treat?
- the risk of recidivism is central, not the disorder
- focus on the risk factors
- RNR would not treat a mental illness if present (and if it is unrelated to re-offending)
3) need principle – what to treat?
- focus on criminogenic needs which are related to future re-offending behavior that has
been predicted (for the specific individual)
- dynamic risk factors (which can be changed) are targeted
- central 8 risk factors according to Andrews & Bonta (first they were the big 4 and then
the moderate 4 were added)
4) responsivity principle – how to treat?
- people with intellectual disabilities should not be treated with CBT
- droping out of treatment is a risk factor
- external responsivity – related to characteristics of the psychologist/ availability of the
treatment possibilities
ex. a young mother psychologist should treat pedophiles
- internal responsivity – individual characteristics of the offender/ personal learning
styles/ motivation of the patient
- responsive treatment is associated with a reduction in recidivism
2.2 the central 8 (criminogenic needs)
history of antisocial behavior (static factor)
- history cannot be changed
- the dynamic need would be used to change antisocial behavioral tendencies
antisocial personality pattern
- characteristics associated with higher risk of offending
antisocial cognitions
- CBT is widely used
antisocial peers
- not having contact with non-criminal others is a risk factor
family/marital relationships
- education
- level of care you have received/ receive
- supervision while growing up
- parenting style
school and/or job
- not having a steady job, dropping out of school and low levels of satisfaction are risk
factors
, pro-social recreational activities
- sports, extracurricular activities
- criminals do not participate in pro-social groups
substance abuse
remorse/ guilt are not established as factors that reduce recidivism
2.3 non-criminogenic needs
focus on the risk principle only leads to loss of a lot of information
dynamic risk factors that are indirectly related to the increase in general offending
- psychiatric disorder
- emotional problems
- self-esteem
- health
Crime Behavior Checklist – looks at the characteristics of the crime scene (not exam
material)
questions:
- Lombroso said that crime is biologically motivated
- the risk principle states that you should take into account who we need to treat
LECTURE 2
1. Predictive validity
is the tool able to predict behaviors/ re-offending?
Singh: it is important to know if the instrument is influenced by the base rate/ do we need
to have a cut-off score?
- low base rate, bad prediction
1.1 area under the curve (AUC)
comes from the receiver operating curve
classifies offenders vs non-offenders
0 to 1 (1 is perfect prediction)
- 0.50 – as good as chance
- 0.71 and above – the instrument is a good predictor
a fairly simple way of assessing risk
- used in cancer measurements
the most commonly used
AUC is a retrospective measure – we look at recidivists’ history of offending
prospective measure – we follow a person in order to see if they’ll re-offend
sensitivity and specificity
- both are retrospective
- sensitivity is about recidivists
- specificity is about non-recidivists (in the slide should be “non-recidivists classified as
low risk”)
advantage of the AUC: no cut-off score is needed
disadvantage of the AUC: it does not give any information about the patient, it just
provides a measure of the validity
AUC is not sensitive to the base rate/ it does not depend on the base rate
1.2 NPV/ PPV
prospective measures
1.3 NSD/NND
, clinically relevant
identification of high-risk patients is always difficult (low-risk patients are always easier
to identify)
2. Generations of risk assessment instruments
first generation – unstructured professional judgment
second generation – actuarial judgment
- no clinical judgment
- static risk factors
third generation – structured clinical judgment
- the clinician uses lists/ risk factors that have been validated by research
- dynamic risk factors
fourth generation – combined with case management
- both static and dynamic risk factors
2.1 unstructured professional judgment
no risk assessment tool is required
2.2 actuarial instruments
a list of factors that have been shown to be related to re-offending
evidence-based instruments
static items are used
ex. age of first offense
advantages:
- quick and easy is not the case for all instruments
disadvantages:
- atheoretical – we do not know why there is a relationship between the factors and
offending
- no clinical override – items that are not related to offending of a person are also added to
the final score of the patient
- difficult to make conclusions about the individual level
the PCL-R is good at predicting violent re-offending
- not a risk assessment tool but used as such
StatRec – actuarial instrument
- only 6 items
- good predictive validity (.80)
Static99-R – actuarial instrument
- 10 static risk factors
- no clinical judgment included
training is required in order to work with actuarial instruments
2.3 structured clinical judgment
combination of actuarial and clinical method
- risk factors could be added/ excluded depending on the clinician’s perspective
mostly dynamic risk factors
used to measure change over time
good predictive validity
professionals decide what is important and what is not
- they can focus on context-dependent factors/ add risk factors to the list that they find
important
includes factors (multi-dimensional)
LECTURE 1
1. Case study – the corn-killing
court psychiatrist assessment:
- introverted (not part of the central 8/ non-criminogenic need)
- tendency to bottle up aggression (not part of the central 8/ non-criminogenic need)
- difficulty empathizing with the feelings of others (part of the central 8)
- no personality disorder (it was not mentioned how the psychiatrist came to that
conclusion/ not part of the central 8)
- anger issues regarding his mother (not part of the central 8)
- very high chances for recidivism unless treatment is provided
! the risk of recidivism should be decided first, and then the need for an intervention
should be established
- no guilt/ no remorse (not part of the central 8)
2. History of risk assessment
older and newer definitions
- they portray different reasons for risk assessment
- the first one points to the severity of the crime – it is not that important
- the second definition – risk assessment is not only determining risk but also identifies the
need of interventions
several definitions are available
risk management – the chance of re-offending is never 0, but what can be done to reduce
the risk factors a person has?
why do we do risk assessment?
- increase safety
- public interest
- ethical importance
- therapeutic importance
- communication
Lombroso (not completely wrong)
- criminal behavior is inborn and it is shown in one’s physical characteristics
- characteristics related to higher chances of becoming violent
father of risk assessment – Burgess
- created the first risk assessment tool
- the tool was actuarial in nature
- re-offending rates dropped
- data from offenders who were on parole
- the instrument has been validated
Baxstrom v Herold (1966) – American case
- Baxstrom was released and many high-risk patients were let back into society or sent to
regular mental institutions (not forensic psychiatric hospitals)
expertise of the expert study
- teachers who had no background in forensic psychiatry/ law
- low IRR between the teachers and the specialists
- many similarities but the teachers did better than the psychiatrists
, - the more information that is available, the more conservative one is in their judgment
- psychiatrists guess correctly 1/3
the 90s
- the Dutch TBS system was praised across the world
- denial/ lack of empathy – risk factors
- the experts were wrong when pointing out important risk factors
Monahan – investigated true risk factors and the RNR model was developed later
2.1 RNR model
4 main principles (22 in total):
1) evidence-based practice
2) risk principle – who to treat?
- the risk of recidivism is central, not the disorder
- focus on the risk factors
- RNR would not treat a mental illness if present (and if it is unrelated to re-offending)
3) need principle – what to treat?
- focus on criminogenic needs which are related to future re-offending behavior that has
been predicted (for the specific individual)
- dynamic risk factors (which can be changed) are targeted
- central 8 risk factors according to Andrews & Bonta (first they were the big 4 and then
the moderate 4 were added)
4) responsivity principle – how to treat?
- people with intellectual disabilities should not be treated with CBT
- droping out of treatment is a risk factor
- external responsivity – related to characteristics of the psychologist/ availability of the
treatment possibilities
ex. a young mother psychologist should treat pedophiles
- internal responsivity – individual characteristics of the offender/ personal learning
styles/ motivation of the patient
- responsive treatment is associated with a reduction in recidivism
2.2 the central 8 (criminogenic needs)
history of antisocial behavior (static factor)
- history cannot be changed
- the dynamic need would be used to change antisocial behavioral tendencies
antisocial personality pattern
- characteristics associated with higher risk of offending
antisocial cognitions
- CBT is widely used
antisocial peers
- not having contact with non-criminal others is a risk factor
family/marital relationships
- education
- level of care you have received/ receive
- supervision while growing up
- parenting style
school and/or job
- not having a steady job, dropping out of school and low levels of satisfaction are risk
factors
, pro-social recreational activities
- sports, extracurricular activities
- criminals do not participate in pro-social groups
substance abuse
remorse/ guilt are not established as factors that reduce recidivism
2.3 non-criminogenic needs
focus on the risk principle only leads to loss of a lot of information
dynamic risk factors that are indirectly related to the increase in general offending
- psychiatric disorder
- emotional problems
- self-esteem
- health
Crime Behavior Checklist – looks at the characteristics of the crime scene (not exam
material)
questions:
- Lombroso said that crime is biologically motivated
- the risk principle states that you should take into account who we need to treat
LECTURE 2
1. Predictive validity
is the tool able to predict behaviors/ re-offending?
Singh: it is important to know if the instrument is influenced by the base rate/ do we need
to have a cut-off score?
- low base rate, bad prediction
1.1 area under the curve (AUC)
comes from the receiver operating curve
classifies offenders vs non-offenders
0 to 1 (1 is perfect prediction)
- 0.50 – as good as chance
- 0.71 and above – the instrument is a good predictor
a fairly simple way of assessing risk
- used in cancer measurements
the most commonly used
AUC is a retrospective measure – we look at recidivists’ history of offending
prospective measure – we follow a person in order to see if they’ll re-offend
sensitivity and specificity
- both are retrospective
- sensitivity is about recidivists
- specificity is about non-recidivists (in the slide should be “non-recidivists classified as
low risk”)
advantage of the AUC: no cut-off score is needed
disadvantage of the AUC: it does not give any information about the patient, it just
provides a measure of the validity
AUC is not sensitive to the base rate/ it does not depend on the base rate
1.2 NPV/ PPV
prospective measures
1.3 NSD/NND
, clinically relevant
identification of high-risk patients is always difficult (low-risk patients are always easier
to identify)
2. Generations of risk assessment instruments
first generation – unstructured professional judgment
second generation – actuarial judgment
- no clinical judgment
- static risk factors
third generation – structured clinical judgment
- the clinician uses lists/ risk factors that have been validated by research
- dynamic risk factors
fourth generation – combined with case management
- both static and dynamic risk factors
2.1 unstructured professional judgment
no risk assessment tool is required
2.2 actuarial instruments
a list of factors that have been shown to be related to re-offending
evidence-based instruments
static items are used
ex. age of first offense
advantages:
- quick and easy is not the case for all instruments
disadvantages:
- atheoretical – we do not know why there is a relationship between the factors and
offending
- no clinical override – items that are not related to offending of a person are also added to
the final score of the patient
- difficult to make conclusions about the individual level
the PCL-R is good at predicting violent re-offending
- not a risk assessment tool but used as such
StatRec – actuarial instrument
- only 6 items
- good predictive validity (.80)
Static99-R – actuarial instrument
- 10 static risk factors
- no clinical judgment included
training is required in order to work with actuarial instruments
2.3 structured clinical judgment
combination of actuarial and clinical method
- risk factors could be added/ excluded depending on the clinician’s perspective
mostly dynamic risk factors
used to measure change over time
good predictive validity
professionals decide what is important and what is not
- they can focus on context-dependent factors/ add risk factors to the list that they find
important
includes factors (multi-dimensional)
