THUR, Oct 7 \ 9.00-11.00 \ Risk Assessment, III

050
REFLECTING ON THE RELIABILITY OF STATIC RISK FACTORS POST- TREATMENT.

Scott A Aylwin, Lea H. Studer, & John R. Reddon
Phoenix Program, Alberta Hospital Edmonton
Lea.studer@amhb.ab.ca

This paper updates recent studies from the Phoenix Program, at Alberta Hospital Edmonton that relate to actuarial risk prediction. Prior number of convictions and relationship to victims (intra- or extra-familial) have been extensively studied and found to be valid predictors of recidivism risk among sex offenders. Despite statistical support of these static risk factors, they may also be misleading if not judiciously applied in a risk assessment algorithm. For example, the highly robust static variable of number of prior offenses, has been shown to be far less predictive of recidivism after our treatment program, than without treatment. We anticipate this to be a reflection of treatment effect. Another example involves the characteristics of the index offense. In our group of offenders against children, there were far fewer differences between groups of offenders, based on the relationship to index victim than is widely accepted. Lastly, a modest positive correlation between testosterone level and sexual recidivism, while demonstrable for treatment non-completers was not significant for completers. Data for this recent finding and clinical implications will be discussed.

051
THE RELATIONSHIP OF STABLE DYNAMIC RISK DOMAINS TO RECONVICTION

Leam Craig, Anthony Beech, Kevin Browne
Forensic Psychology Practice Ltd and University of Birmingham, UK.
leamcraig@forensicpsychology17.freeserve.co.uk

The aim of this study was to cross validate the predictive accuracy of the Thornton's Deviancy Domains using measures of psychosexual characteristics and psychopathology. The Special Hospitals Assessment of Personality and the Multiphasic Sex Inventory (MSI) were administered to 119 sexual offenders. Scales measuring one of the four Deviancy Domains, identified by Thornton (2002), sexual interests, distorted thinking, socio-affective functioning and self management. The accuracy of the domains, in predicting sexual reconviction over two and five year follow-up periods, was measured using the area under the curve analysis. Sexual Interests and Distorted Thinking domains were found to be significantly correlated with sexual reconviction over the two and five year period. The Sexual Interests domain in particular obtained moderate accuracy in predicting sexual reconviction (AUC =. 86 over 2 years) and outperformed actuarial risk predictions (Static-99) (AUC =. 66 over 2 years). The results support the use of identifying dynamic risk domains and their utility in predicting sexual reconviction.

052
PUPILLOMETRY: A POSSIBILITY TO DIFFERENTIATE PEDOPHILES FROM NON-PEDOPHILES

Frottier, Kaplan, Linzmayer, Gruenberger
Federal Documentation Centre for Sexual Offenders, Vienna, Austria
patrick.frottier@justiz.gv.at

Psychosensory stimuli (e.g.: emotion, vision) stimulate the sympathetic system, while inhibiting the parasympathetic system. Pupillary reflex dilatation is based on the reciprocal influences of the sympathetic and parasympathetic system.
The pupil dilates as a result of this reciprocal interaction.
Since the beginning of the seventies different investigations have tried to prove, that the use of reflex pupillary dilation is an objective indicator for discriminating pedophiles from non-pedophiles. Discrepant findings stopped further research, although it was probable, that insufficient methodology had been the reason.
Research in the eighties and nineties have developed sophisticated methodologies, that are regularly used in clinical settings. The leading researcher in this field, the Austrian J. Gruenberger (who is also directing this research project) developed a technique for pupillometric measurements utilizing different static and dynamic variables after presenting various visual stimuli to the patient for differentiating subtypes of alcohol dependency.
Recently researches in the USA and in Austria have independently started to reinvestigate the relationship between psychosensory stimuli and pupillary dynamics to discriminate sexual offenders and especially pedophiles from non-pedophiles.
Pupillography was used to examine the stimulation caused by pictures of prepubescent boys and girls to 3 different groups: pedophile sexual offenders, non-pedophile sexual offenders and normal controls.
This presentation will focus on the methodology and the very encouraging results of the Austrian research.