Category: Criminal Justice / Forensics

PS9- #B33 - Risk Factors for Paraphilic Sex Offending in Japan

Saturday, Nov 18
9:45 AM – 10:45 AM
Location: Indigo Ballroom CDGH, Level 2, Indigo Level

Keywords: Risk / Vulnerability Factors | Relapse | Aggression / Disruptive Behaviors / Conduct Problems

Manifestations of paraphilic sex offending considerably vary across countries because sexual behaviors are widely influenced by social and cultural factors. In Japan, frotteurism and voyeurism are among the most major sexual offending but these behaviors are rarely seen in other countries. More than 4,000 chikans (male frotteurists) are arrested each year for offences carried out on public transport. The aim of this study is to identify risk factors associated with recidivism of paraphilic sex offending.


Participants were 137 treatment-seeking male paraphilic offenders. All participants were treated in a private psychiatric hospital for 6 months and were observed for recidivism for one year after treatment. A total of 11 participants recidivated during the follow-up period. The demographics of those relapsed (N=11) and abstinent (N=126) were compared. As a result, there was a significant difference in co-morbidity (54.6% vs 5.7%)(χ2(2)=21.03, p < 0.01), the age of onset (17.8±4.9 years vs 24.7±8.0 years) (t(136)=-3.99, p < 0.01), and the Static-99 score (5.0±1.6 vs 3.3±1.5) (t(136)=3.26, p < 0.01). Also, recidivated participants had more previous criminal records at the marginally significant level (χ2(1)=3.56, p < 0.1).


These variables (co-morbidity, onset age, number of arrest) were then included in a forward stepwise logistic regression model. Since the age of onset and the criminal record are also the items included in the Static-99, its score was not included in the model to avoid multicollinearity. The AUC was 0.92 and Pearson’s goodness-of hit test indicated that the model fit well (p = 0.65).It was found that 47.3% of recidivism could be explained by these three variables. The most powerful variable to explain recidivism was co-morbidity (β=3.76, p=0.01, 95%CI=0.77 to 4.40), followed by the number of arrest (β=1.12, p=0.02, 95%CI=0.21 to 2.03) and the age of onset (β=-0.21, p=0.04, 95%CI=-0.43 to -0.01). The marginal effect of the onset age was calculated and the result indicated that the probability of relapse increases by 2% (p=0.02, 95%CI=-0.035 to -0.030) if onset gets earlier by one year.

It is found that these risk factors are basically the same as those found in previous studies conducted in Western countries. The results can be used for the screening purpose and are of help to inform  patient-treatment matching.

Takayuki Harada

Professor
University of Tsukuba

Masayoshi Hojo

Enomoto Clinic