ETHICAL ISSUES IN RISK ASSESSMENTS
July 27, 2009
RISK - DEFINITION
• Risk is a “threat or hazard that isn’t completely understood and therefore can be forecast only with uncertainty.”
• Risk involves the ideas of “nature, severity, frequency, imminence, and likelihood – not just probability.”
definition
• Risk is “context specific”
• Risk is “never known but estimated.”
RISK ASSESSMENT IS:
• “Gathering information to assist in decision-making”.
• “An individualized process”
RISK ASSESSMENT ISN’T:
• “Simply providing a diagnosis or prognosis”
• “Simply considering a set of test items or risk factors determined before the evaluation.”
GOALS OF RISK ASSESSMENT
• To contain and reduce the individual’s risk.
• To guide interventions.
• To improve the consistency of decisions.
• To improve the transparency of decisions to protect the rights of the individual, as well as the community and potential victims.
ESTIMATING FIRST OFFENSE
• Intent
• Plan
• Means
• Opportunity
• Coin Toss
ESTIMATING RECIDIVISM
• History (static factors)
• Dynamic Factors
• Acute Factors
Your Perspective
LEVEL # % VOR
Low 900 30% 270
High 100 90% 90
DETECTION RATES
FREQ VICTIMS / 20 YRS
5 YRS 15 YRS 20 YRS
LOW 4 18.5% 47.8% 68.4%
HIGH 20 64.1% 96.1 99.7%
WHO IS MORE DANGEROUS?
• Adult male sex offender
– Selects young boy victims
– Selects young girls victims
RESEARCH CRITERION
• Re-arrest / hospitalization for anything
• Re-arrest / hospitalization for violence
• Charged for anything or new offense
• Convicted for anything or new offense
VICTIMIZATION STUDIES
Besserer and Trainor (2000)
• 78% of sexual assaults not reported to anyone.
• 59% “not important enough”
POLICE RECORDS
• 50% of sexual assault victims will report
• Incest victims unlikely to report first offense and almost never report the second
• Previous victims who went through the system unlikely to report a new offense
CLINICAL RECORDS
• Diagnoses to justify services
• Lack information about crimes
• Confidentiality
• Different providers
HOW CHARGED?
• Plea bargain = 25 victims become 1
• Specific offense may not be prosecuted
LENGTH OF FOLLOW-UP
• Highest number within first five years after release from prison
• In reality risk period may be 20 – 25 years
• Current studies 6 months to sixteen years
DEVIATIONS FROM GROUP
• For sex offending, the risk of sexual assaults increases to the age of 26 and then decreases to the age of 40, and it is almost nonexistent after the age of 60.
PA AGES AT CONVICTION(n = 6406)
• 63.7% < 40 years old
• 31 % = 41 – 60 years old
• 5.6 % > 60 (7 > 80 years)
SURVIVOR ANALYSIS
• 23% of the individuals receiving the same score as John on the Knapp Risk Assessment Protocol (KRAP) were re-arrested in 2 years, 34% were re-arrested in 5 years, and 43% were re-arrested in 10 years.
HOW DO WE REPORT RISK?
• Risk prediction (dichotomous)
• Risk Estimate (Survival analysis)
• Risk Statement
MENTAL HEALTH ISSUES
• There is no Axis 1 diagnosis with a significant relationship to violence.
• Iterative Classification Tree (Monahan)
SUBSTANCE ABUSE ISSUES
• May be related to first offense
• Recidivists more likely to have co-occurring Antisocial Personality Disorder
• Substance use may be a means, rather than the end.
FRONTAL LOBE ISSUES
• Frontal lobe impairment may increase risk for getting caught, rather than for violence.
THE PRIMARY QUESTIONS?
• To what degree does a person make a choice to be violent?
• To what degree does he or she not make a choice?
• If he or she does not make a choice, what do we do?