Transcript 2014 SOMMI Presentation - Sex Offender Recidivism Commission

Sharon Kelley, Psy.D.
Sand Ridge Evaluation Unit
Madison,WI

Sexual offenders with Major Mental Illness
(SOMMI) are often underserved
◦ Traditional mental health system lacks expertise in the
management of sexual deviance
◦ Traditional sex offense-specific treatment programs
often do not consider unique psychiatric issues


“Best practices” is based on research samples
that do not include SOMMI
Tension between what is known/available for this
population and risk/treatment needs of
population
Study
Sample Type
MMI Rate
Cochrane et al. (2001)
Court clinics across
U.S.
N=1,710
Psychotic d/os less
freq among sex
offenders (16%)
compared to general
offenders (32%)
Becker et al. (2003)
120 sex offenders
awaiting trial for civil
commitment for SVP in
AZ
50% of the Axis I d/os
identified were rel to
paraphilias and
substance use
Langstrom et al.
(2004)
1,215 convicted sex
offenders in Sweden
-34.4% had a psych
hosp at some point
-1.4% met criteria for
a psychotic disorder
Study
Sample Type
Offense Rate
Wallace et al. (2004)
2,861 patients with
Schizophrenia in
Australia over a 25year period
Sex offense
convictions = 1.8%
Fisher et al. (Mass
Mental Health –
Criminal Justice
Cohort Study, 2006)
Arrest records during
a 9 yr period for all
DMH clts
(N = 13,978)
Of the 17,000 arrests,
only 272 (1.6%) were
for sex offenses
But 255/272 offenses
were for serious
charges: Indecent
Exposure, Indecent
A&B on an adult and
child, Rape


Very few studies
Hanson & Bussiere (1998) meta-analysis
◦ “The large correlation for our ‘severely disordered’
variable could be almost completely attributed to
Hackett’s (1971) report that all of his exhibitionists
with psychotic symptoms eventually recidivated” (p.
353).
◦ Association not found in follow-up meta-analysis
(Hanson & Morton-Bourgon, 2004)

Hanson et al. (2007; DSP)
◦ Twice as many MMI subjects re-offended (18%) as
compared to the total sample (9%)

Inconsistent findings regarding the relationship
between MMI and risk for violence in the general
pop
◦ Bonta et al. (1998) found that the average association
between psychosis and violence was small and negative
(r=.04) across the 11 studies in their meta-analysis
reporting on psychosis (results limited to MDOs released
from a correctional setting)
◦ Douglas et al. (2009) meta-analysis involving 204
studies: psychosis associated with a 49%-68% increased
likelihood for violence.
 Effect size depends on presence of moderators but MMI
found to be a strong risk factor for violence compared to
persons without MMI


The relationship b/n violence and MMI is
strengthened by the presence of other risk
factors: psychopathy, ASPD, substance use
(Douglas et al., 2009; Fisher et al., 2006;
Monahan et al., 2001)
Positive symptoms of psychosis more strongly
related to violence (Douglas et al., 2009)
◦ Swanson et al. (2006) reported on 1,410 patients with
schizophrenia drawn from 57 mental health sites across
24 states. They found that positive symptoms of
schizophrenia were associated with both minor and
serious violence, even after controlling for numerous
possible confounds and covariates.



No comprehensive study like MacArthur Risk
Assessment Study for SOMMI
Very broadband definitions used in SO
studies, such as “any mental disorder,” are
apt to blur important distinctions between
specific psychotic syndromes (Douglas et al.,
2009)
Most articles are descriptive with small
sample sizes
◦ Still provide a good start…

Phillips et al. (1999) 17 pts with Schizophrenia
◦ Sex offending usually postdated onset of psychosis
◦ Majority were psychotic at the time of the offense
◦ Psychosis was not a direct causal factor but contributed
to disinhibited sexual behavior.

Craissaiti & Hodes (1992) 11 pts with
psychosis
◦
◦
◦
◦
◦
Sex offenses generally non-violent and impulsive
No evidence the pts attempted to evade capture
Victims mostly adults and known to offender
Only 1 pt taken admitted to a hosp following arrest
4 pts had engaged in mast fantasy prior to offending

Phillips et al. (1999) 15 SOs with Schiz
◦ None had a hx of sexual promiscuity
◦ Little to no hx of long-term intimate partners
◦ Compared to MMI without hx of SO, this group was
twice as likely to report an unimpaired sexual
interest

Sahota & Chesterman (1998) 20 SOMMI pts
◦ None had a stable intimate relationship lasting
longer than 12 weeks
◦ Psychotic break usually occurs at a crucial age
period when many pts are dev a sexual identity and
establishing intimate sexual relationships

Greenall & Jellicoe-Jones (2007) 11 cases
◦ 3 subjects were psychotic at the time and offenses
driven by anger that was exacerbated by psychosis
◦ 4 cases primarily driven by psychosis
◦ 2 cases were sexually inhibited
◦ 2 cases had underlying paraphilias
◦ Concluded: the presence of MMI may exacerbate
risk factors by reducing effective self-regulation

Smith & Taylor (1999)  84 pts with Schiz
hosp after conviction for a sex offense
◦ 80 pts committed offenses when actively psychotic
◦ 4 pts had onset of psychosis following offense
Direct
Indirect
Coincidental
Not
present
Total
%
Delusion
s
18%
25%
51%
6%
N=80
%
Hallucinations
15%
18%
45%
22%
N=80
Relationship Between SO and
MMI

Factors to Consider:
◦ Onset of MI sxs in relation to onset of
PSBs
◦ How do PSBs manifest or change when
psychiatrically decompensated?
◦ Are PSBs present when psychiatrically
stable?
◦ How is PSB manifested in this MI
individual?

How Relationship Impacts
Risk
Increased MI Sxs =
Increased Impulsivity
Increased Hypersexuality
Decreased Behavioral Controls
Decreased Ability to Consider Consequences
Decreased Ability to Make Rational Decisions
Decreased Ability to Engage in Treatment in a
Meaningful Way
◦ Decreased Ability to Plan and Influence Others
◦ Complicated relationship b/n PI and Grievance
Thinking
◦ Impaired social and intimacy skills
◦
◦
◦
◦
◦
◦






27-year-old male
Paranoid Schizophrenic Disorder
Risperdal, Benadryl, Ativan, and Aterax
Charged with 3 counts Indecent A&B < 14 but
found NGI
Had not been med compliant for several months
prior to his offenses
Offenses related to a fixed delusion that people
were actually robots, that he had magical powers,
and that he could have “eye sex” with children






47-year-old male released from prison out of state
after serving 15 years for 3 counts of Criminal
Sexual Conduct, 1 count Gross Indecency, and 1
count Assault with Intent to Commit Criminal
Sexual Conduct
Victims were all under-aged boys
Self-reported approximately 100 victims for which
he did not get caught
Recently re-arrested for possession of child porn
Dx: Bipolar Disorder and Pedophilic Disorder
Although mania increases hypersexuality, evidence
suggests it didn’t play a significant role




Paranoid Schizophrenic Disorder
Prolixin, and Cogentin
Convicted of Rape and 2 counts Indecent
A&B
When non compliant with meds has
engaged in more primitive type of offenses
(i.e., leering at women, following women,
exposing self). When compliant with meds
has engaged in more well organized
offenses such as grooming behaviors,
getting victim incapacitated (i.e., alcohol),
kidnapping, and rape.




Same case
Less meds = more disinhibited, impulsive,
disorganized sexual behavior
As he gets more psychotic, though, the
problematic sexual behaviors decrease
When acutely psychotic, he is catatonic =
Absence of Risk.

Langstrom et al. (2004) N=1,215 sex
offenders
◦ 4% had a psych hosp in within the year preceding
the index offense
◦ Sexual recidivism was found to be associated with
psychosis, any psychiatric disorder, and any
inpatient care.
◦ However, a prior diagnosis of etoh abuse/dep more
than doubled the odds of a sexual reconviction
◦ A personality disorder diagnosis increased the odds
by a magnitude of ten times

Olver, Stockdale, & Wormith, 2011 Meta-analysis
of attrition (included sex offender, domestic
violence, general correctional, and violent
nonsexual offender programs)
◦ Major mental illness was a mediating variable for
recidivism
◦ MMI pop (psychotic disorders and BPDs) less likely to
complete treatment across all programs. Those who
were less likely to complete treatment were more likely
to recidivate.
◦ However, MMI was not the strongest correlate.
 Young, single, unemployed, ethnic minority, male, limited
formal ed, low SES, hx of prev offenses, high static risk


Abracen & Looman (2012 ATSA Conference):
Examined 348 high risk sex offenders. Found
that after controlling for risk scores on the
Static-99R, only those with a history of
psychiatric impairment was found to add
incrementally to predict recidivism.


Only study to date to specifically examine the
predictive validity of a dynamic risk measure
in the SOMMI population
Problems:
◦ Small N = 61
◦ Coded by probation officers
◦ Major mental illness defined as: at least one night
in a hospital
◦ Unknown whether this was due to depression,
bereavement, adjustment disorder, personality
disorder, malingering, etc…..?
Major mental disorder
Recidivists
18% (11/61)
Static-99R
.744*
Static-2002R
.727*
STABLE-2007
.595
Static-99R/STABLE-2007
.669
Static-2002R/STABLE-2007
.709*
1.
2.
3.
4.
Are there risk factors that are unique to the
SOMMI population?
Do criminogenic needs operate independently
of symptoms of a major mental illness?
Does the presence of a major mental illness
exacerbate pre-existing criminogenic needs?
Does the presence of a major mental illness act
as a protective factor and serves to moderate
the effect of pre-existing criminogenic needs?

1
2
3
4
5
Specific Aims:
Determine whether the SRA-FV can be scored
reliably on the SOMMI population.
Develop supplementary scoring guidance and
training to facilitate the reliable application of
the SRA-FV to the SOMMI population.
Develop SOMMI norms for the SRA-FV.
Identify groupings of criminogenic needs within
the SOMMI population (and whether they have
needs that are unique to this population).
Explore whether acute symptoms have a
moderating or worsening effect on existing
criminogenic needs.