ORIGINAL RESEARCH

published: 03 March 2020

doi: 10.3389/fpsyt.2020.00088

Prevention of Sexual Child Abuse:

Preliminary Results From an
Outpatient Therapy Program
Tamara S. N. Wild 1,2†, Isabel Müller 2†, Peter Fromberger 1, Kirsten Jordan 1, Lenka Klein 2
and Jürgen L. Müller 1,2*
1Clinic for Psychiatry and Psychotherapy – Forensic Psychiatry, Human Medical Center Göttingen, Georg-August-University
Göttingen, Göttingen, Germany, 2 Prevention of Sexual Abuse (PsM), Asklepios Psychiatric Clinic, Göttingen, Germany
Edited by:
Athanassios Douzenis,
National and Kapodistrian University of In Germany, access to outpatient treatment services devoted to the prevention of (further)
Athens, Greece
sexual offenses against minors and child sexual exploitation material (CSEM) offenses is often
Reviewed by:
limited. The therapy project “Prevention of Sexual Abuse” tries to fill this gap by providing
Kolja Schiltz,
Ludwig Maximilian University of treatment to patients with a self-reported sexual interest in children and adolescents,
Munich, Germany irrespective of whether or not they are pedophilic or prosecuted by the legal justice system.
Bernhard Bogerts,
Otto von Guericke University
Within the project, a treatment manual was developed which specifically addresses dynamic
Magdeburg, Germany risk-factors in child sexual abusers and CSEM offenders. The treatment manual was conceived
*Correspondence: to reduce recidivism risk and to contribute to the enhancement of the patients’ personal well-
Jürgen L. Müller
being. In this paper, results of the accompanying scientific research are presented: offense-
[email protected]
†
supportive attitudes (N = 23), self-reported CSEM use (N = 10), emotional distress (N = 24), and
These authors have contributed
equally to this work participants’ subjective risk perception of committing (further) sexual offenses (N = 25) reduced
during the course of treatment. A reduction of offense-supportive attitudes was further
Specialty section: observed from pre-intervention to 1-year follow-up (N = 8). Changes with regard to self-
This article was submitted to
Forensic Psychiatry, efficacy, quality of life, participants’ self-perceived ability to control sexual impulses toward
a section of the journal children and adolescents permanently, and several measures assessing different kinds of
Frontiers in Psychiatry
sexual recidivism did not, however, reach any level of significance. During an average
Received: 17 October 2019
Accepted: 03 February 2020
observation period of 2.4 years, six patients confessed to have conducted new sexual
Published: 03 March 2020 exploitation material offenses, while no further sexual abuse cases were reported (N = 19).
Citation: Due to the used research design and small sample sizes, treatment effects cannot be inferred
Wild TSN, Müller I, Fromberger P, and external validity is limited. This notwithstanding, results provide first evidence for a
Jordan K, Klein L and Müller JL (2020)
Prevention of Sexual Child Abuse: relationship between treatment participation and self-reported recidivism and psychological
Preliminary Results From an well-being.
Outpatient Therapy Program.
Front. Psychiatry 11:88. Keywords: child sexual abuse, child sexual exploitation material, child pornography, sex offender, pedophilia,
doi: 10.3389/fpsyt.2020.00088 treatment, therapy, well-being

Frontiers in Psychiatry | www.frontiersin.org 1 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

INTRODUCTION sexual orientation toward children (12). Accordingly, offenders

should be provided with treatment irrespective of potential
In recent years, media attention has increasingly focused on child pedophilic interests.
sexual abuse (CSA), hence raising the public awareness toward Cognitive behavioral therapy programs applying the risk-need-
the need and importance of prevention programs for this specific responsivity principles and addressing dynamic risk factors have
offender group. In the literature, many definitions have been been shown to be most effective in the treatment of sex offenders
proposed for CSA (1, 2). What they have in common is that CSA (13, 14) and are of importance for a number of reasons: (1) having
does not need to include physical contact between a perpetrator persistent sexual urges involving children can be experienced as
and a child. Instead, the definitions refer to different kinds of markedly distressing and may therefore require treatment (15), (2)
sexual harassment that can occur on a continuum of power and the committal of sexual offenses against children can entail a host
control, ranging from non-contact sexual assault (e.g., of serious penalties, including substantial fines, probation, or jail
exhibitionistic acts) to contact sexual assault (e.g., forcible sentences and can also result in the loss of significant others or
rape). Moreover, the definitions also include internet sexual social exclusion. (3) CSA is linked to a number of adverse
offending, that is, the exploitation of children online. As a consequences for the affected children. Indeed, minors who have
consequence of the advent of new digital technologies and the been abused sexually may develop a variety of mental health
growth of the Internet, possibilities to commit offenses against problems such as affective disorders, suicidal behavior, alcohol,
the sexual self-determination of minors from behind the drug and medication dependence, social anxiety, conduct
computer screen have increased. Both the illicit distribution, disorder, borderline personality disorder, posttraumatic stress
acquisition, and possession of child sexual exploitation material disorder, eating disorders, especially bulimia nervosa, or an
(CSEM) and online grooming and solicitation, the initiation of increased risk of revictimization (16–19). Moreover, there is
online contacts with children with the intention of gratifying evidence that children whose sexual abuse has been recorded
one’s sexual desire by means of the receipt of sexually explicit and distributed on the internet additionally suffer once they realize
images or cybersex (3) fall in that category. Research indicates that their indecent images cannot be removed from the Internet
that the number of internet sexual offenses has increased (4, 5), and that they are continuously being victimized by a large number
aided by a phenomenon that can be traced to the ease of of offenders (20, 21). In summary, the consequences of CSA and
accessibility at affordable costs, while feeling secure due to the CSEM offenses are detrimental for both the offender and the
anonymity of the online environment [the so-called Triple A victim, which emphasizes the importance of out-patient
Engine: accessibility, affordability, and anonymity (6)]. prevention programs.
Prevalence rates of sexual offending against minors are
difficult to estimate. Data from both official arrest statistics and The Outpatient Treatment Facility
self-report studies may result in under-reporting. Different “Prevention of Sexual Abuse” (PsM)
research groups have nevertheless attempted to examine the Originally, the provision of treatment for CSAs and child sexual
prevalence of CSA and CSEM offenses. For instance, Alanko exploitation material offenders (CSEMOs) in Germany was
et al. (7) were able to show that within a large sample of 3,909 allocated to correctional institutions as well as mental health
Finnish men between the age of 21 and 43, 0.3% indicated to care services. However, while sex offenders who are sentenced to
have had sexual contact with a person under the age of 16. In an more than 2 years in prison receive mandatory treatment in
online study with 8,718 participants (8), exclusive consumption correctional institutions, access to outpatient treatment services
of CSEM for sexual gratification was reported by 1.7% of was often limited. The reasons for this are diverse and range from
subjects, exclusive CSA by 0.8%, and both CSA and CSEM reservations regarding the patient group, fear of reputational
offenses by 0.7%. Ten percent of the participants reported damage, and a lack of willingness to cooperate with legal
having any kind of online contact with minors, while 5.3% authorities [for an overview, see (22, 23); Brand, 2006, as cited
indicated they have had sexual online contact with minors in (22)]. By virtue of the limited access for the highly stigmatized
(predominantly adolescents). The results further indicate that offender group, in the last decade, a small yet growing number of
one third of sexual online contacts resulted in sexual offline specialized community programs targeting the prevention of
meetings; however, also nonsexual online contacts sometimes (repeated) sexual assaults against minors were established
resulted in sexual offline meetings (9). throughout Germany (24, 25), one of them being the
A sexual interest in children is considered a risk factor for outpatient treatment facility “Prevention of Sexual Abuse”1
both the onset and progression of CSA (10). A sexual preference (PsM) in Göttingen (26). The PsM, which was established in
for children, usually of prepubertal or early pubertal age 2011, addresses both men and women who are concerned about
accompanied by persistent sexual fantasies and urges involving their sexual fantasies and behaviors toward children and
children over a period of at least 6 months, on which the adolescents, irrespective of whether they have already
individual has acted or which causes marked distress or committed an offense against the sexual self- determination of
interpersonal difficulty is described as a “pedophilia” according children. In comparison to other specialized treatment centers, it
to ICD-10 (11). Nevertheless, a sexual interest in minors is is further irrelevant if clients fulfill the diagnostic criteria for
neither a necessary nor a sufficient precondition for offenses pedophilia or are being prosecuted criminally. While voluntary
against the sexual self-determination of children. In contrast, it is
1
estimated that only 50% of child sexual abusers (CSAs) have a https://www.asklepios.com/goettingen/experten/schwerpunkte/psm/

Frontiers in Psychiatry | www.frontiersin.org 2 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

participation, intrinsic motivation, willingness to change, and a informed consent was obtained from all participants. Depending
high self-reported degree of psychological strain are mandatory on the therapy form, group size, and individual characteristics
inclusion criteria for (potential) CSAs and CSEMOs, offenders such as engagement in treatment (e.g., as indicated by homework
with probation conditions can nevertheless commence compliance) or intellectual abilities, treatment length varied
treatment. However, probation conditions cannot be met by between several months and 2 years. Due to changes in data
participating in the program. The treatment program is funded collection methodology, not all participants filled in every
by the State Government of Lower Saxony, the Human Medical questionnaire and not all questionnaires were assessed at all
Center Göttingen, and Asklepios Psychiatric Clinic Göttingen. four points in time [pre-intervention (baseline, T1), after the first
From July 2011 up to August 2019, 340 individuals have half of the treatment manual had been completed (Ti; please note
contacted the therapy project PsM. These callers included legal that measurements at this time point were not included in the
authorities, relatives, medical clinicians and psychotherapists, analyses due to low case numbers), post-intervention (T2), and
and others (e.g., legal guardians or priests). However, the at 1-year follow-up (T3)]. As some participants just finished the
majority of patients initiated contact with the PsM by treatment program recently or have dropped out after T2, little
themselves. In most cases, first contact was preceded by a follow-up data is available. As a consequence, different analyses
house search and many of those concerned reported that they were conducted for participants with (a) pre-, post-, and follow-
had been rejected by other specialized treatment programs up intervention data and (b) pre- and post-intervention data.
because of this. In total, 122 patients started the diagnostic Accordingly, the two different samples will be described
phase. From these patients, 83 have gone through the separately in the following two subsections.
diagnostic phase, while seven patients still participate in it
(current as of September 2019). Almost all patients were of Sample description for participants with pre-, post-,
male sex (n = 121), with a mean age of 37 years (SD = 11.9; range
and follow-up intervention data (sample (a); N = 9)
18–77 years). Out of the 122 patients who were offered to start
Nine men aged 25 to 71 years, with a mean age of 44 years (SD =
the diagnostic phase, 93 were involved in the legal justice system,
13.9) and a mean IQ of 100 [range 80–123; SD = 14.8; (29, 30)]
13 were undetected offenders, 14 dealt with sexual fantasies with
had filled in questionnaires at pre- and post-intervention and 1
minors, but had not yet committed a crime, and two suffered
year after treatment completion. Two had sexually abused a
from pedophilia-themed obsessive–compulsive disorder.
child, three had consumed CSEM, and five had committed both
Interestingly, the proportion of fathers is higher among CSAs
offense types. The majority of patients were involved in the
and mixed offenders compared to CSEMOs (39%, 50%, and 29%,
justice system (n = 8), only one subject was an undetected
respectively). This finding is consistent with previous reports
offender. None of the patients were pedophilic according to
(27) and emphasizes the importance of treatment programs in
ICD-10 criteria (missing values n = 1). However, the majority of
order to protect at-risk children from sexual exploitation.
patients (n = 5) fulfilled criteria for at least one psychiatric
A detailed description of the treatment program based on the
disorder. Three patients fulfilled criteria for affective disorders
first German treatment manual specifically addressing
(F30-F39) and two patients were diagnosed with disorders of
(potential) CSAs and CSEMOs (28) and first results can be
adult personality and behavior (F60-F69), one patient with
found elsewhere (24, 26). In the following, we will present
mental and behavioral disorders due to psychoactive substance
updated results from the ongoing accompanying scientific
use (F10-F19), and another patient with neurotic, stress-related
research as well as data on self-reported recidivism rates. In
and somatoform disorders (F40-F49). Due to differences in
line with previous findings (24, 26) we expect improvements
personal backgrounds, participants either received individual
with regard to (1) general self-efficacy, (2) offense-supportive
or group therapy or both (n = 2, n = 5, and n = 2,
attitudes, (3) self-perceived overall emotional distress, (4) life
respectively). Whenever necessary, additional sessions were
satisfaction, (5) self-perceived ability to control their sexual
offered to patients, meaning that treatment was not
impulses toward children and adolescents permanently, and
fully standardized.
(6) subjective risk perception of committing sexual offenses
from pre-intervention to post-intervention and to 1-year
follow-up. Additionally, we expect a reduction of the frequency Sample description for participants with pre-and
of (7a) child and adolescent sexual abuse, and (7b) the post-intervention data (sample (b); N = 25)
consumption of child and adolescent sexual exploitation The sample included 25 men aged 24 to 71 years, with a mean
material use during the course of therapy. age of 41 years (SD = 11.8) and a mean IQ of 97 [range: 65-123;
SD = 16.3; (29, 30)], who had sexually abused a child (n = 6),
consumed CSEM (n = 12), committed both offense types (n = 6),
METHOD or had not yet conducted any sexual offenses against children,
but were afraid they might do so in the future (n = 1). The
Participants majority of patients were involved in the justice system (n = 22),
Participants were patients from our treatment facility who a smaller proportion were undetected offenders (n = 2) or non-
completed the whole treatment program and volunteered to offenders (n = 1). The majority of patients (n = 22) were not
take part in the study. The ethics committee of the Medical pedophilic based on ICD-10 criteria. However, most fulfilled
University Center Göttingen issued a positive vote and written criteria for at least one psychiatric disorder. After consideration

Frontiers in Psychiatry | www.frontiersin.org 3 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

of various factors such as intellectual abilities, work schedule, and Life Satisfaction Questionnaire
comorbid disorders, participants either received individual or The Life Satisfaction Questionnaire [FLZ (36)] measures general
group therapy or both (n = 5, n = 14, and n = 6, respectively). life satisfaction as well as satisfaction with health status, job,
Tables 1 and 2 in the Supplementary Material provide an income, leisure time, partnership, relationship with one’s
overview of demographic characteristics and psychiatric children, oneself, sexuality, friends and relatives, and housing.
diagnoses of sample (a) and sample (b) grouped by offender The ten different domains consist of seven items each, which
status (involved in the justice system, undetected offenders, non- results in 70 items that have to be rated on a seven-point Likert
offenders) and offender type (CSAs, CSEMOs, individuals with scale (1 = “very unsatisfied” to 7 = “very satisfied”). General life
both offenses), respectively. satisfaction is calculated as the sum of the seven subscales health
status, income, leisure time, oneself, sexuality, friends and
relatives, and housing and range between 49 and 343. Higher
MATERIALS total values indicate a greater general life satisfaction, while
higher subscale values reflect greater satisfaction in the specific
Self-Efficacy
domains. Psychometric properties were shown to be good.
Aachen Self-Efficacy Questionnaire
Validity has been demonstrated by factor analysis and internal
The Aachen Self-Efficacy Questionnaire [ASF (31)] measures
consistency of the different subscales varies between Cronbach’s
generalized self-efficacy as well as self-efficacy for achievement,
a = .82 and a = .95 (36).
social interactions and health-related behaviors. Twenty items
(e.g., “I can trust my abilities”) have to be rated on a five-point
Likert scale, ranging from 1 (“does not apply at all”) to 5 (“fully
Subjective Sexual Self-Regulation
High Risk Situation Test
applies”). Total values can range between 20 and 100, with higher
Subjective risk perception in a variety of situations (e.g., when
values reflecting greater subjective self-efficacy. Psychometric
alone with a child) was assessed by means of the High Risk
properties have been shown to be good. The internal
Situation Test [HRST (37); German version (38)]. The
consistency for the general scale is Cronbach’s a = .90, for the
questionnaire consists of 58 items that need to be rated on a
three subscales it is slightly lower (Cronbach’s a = .74-.84). Over
five-point Likert scale, ranging from “low” to “extremely high”.
a period of 8 weeks, test-retest reliability was rtt = .66 (31).
The total value can vary between 58 and 290, with higher values
representing an increased self-perceived risk to commit
Offense-Supportive Attitudes
sexual offenses.
Bumby Molest Scale
Offense-supportive attitudes, that is, beliefs that excuse or justify
Coping Self-Efficacy Scale Related to Minors—
sexual harassment, were measured with the Bumby Molest Scale
Coping
[BMS (32); in the German version (33)]. The questionnaire
The Coping subscale of the Self-Efficacy Scale Related to Minors
consists of 38 items, an example being “Some sexual
[SESM-C (38)], consisting of 20 items, was used to measure the
relationships with children are a lot like adult sexual
participants’ self-perceived ability to control their sexual
relationships”. Items are rated on a four-point Likert scale,
impulses permanently. On a four-point Likert scale ranging
ranging from “strongly disagree” to “strongly agree”. The total
from “not true” to “absolutely true”, participants have to
value can vary between 38 and 152, with higher values
indicate how certain they feel that they are able to control their
representing stronger offense-supportive attitudes. In the
sexual urges toward children or adolescents permanently in a
original study, the Bumby Molest Scale showed good
variety of situations (e.g., when alone with a child). Lower scores
psychometric properties (Cronbach’s a = .97, rtt = .84).
represent greater deficits in the perceived ability to maintain self-
control. Internal reliability was shown to be high (Cronbach’s
Personal Well-Being
a = .94).
Symptom Checklist-90-Revised
To assess subjective symptoms and psychopathologic features,
the Sympton Checklist-90-Revised [SCL-90_R (34); in the Self-Reported Sexual Offenses Against
German version (35)], a self-report inventory comprised of 90 Children and Adolescents
items on nine subscales (somatization, obsessive-compulsive, In order to assess sexual offenses against minors, two self-report
interpersonal sensitivity, depression, anxiety, hostility, phobic instruments were used—the Sexual Behavior Involving Minors
anxiety, paranoid ideation, and psychoticism) was used. All Scale (SBIMS) (38) and the Sexual Fantasies and Behaviors
items have to be rated on a five-point Likert scale, ranging Questionnaire (SPV) (26). The SBIMS was in use until
from 1 (“not at all”) to 4 (“extremely”). The Global Severity February 2014 and was thereafter replaced by the SPV. Both
Index can be calculated to indicate overall emotional distress, instruments intend to measure the frequency of sexual contacts
with higher total values reflecting a greater subjective burden. with minors and the consumption of sexual exploitation material
Psychometric evaluations have reported adequate internal depicting children and adolescents during the six months
consistency (Cronbach’s a = .79 to .89), and acceptable to preceding the assessment. Furthermore, they both include
good test-retest reliability (35). items concerning the frequency of the occurrence of sexual

Frontiers in Psychiatry | www.frontiersin.org 4 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

fantasies including minors, which were not, however, used for RESULTS
the purpose of this paper.
Self-Efficacy
Sexual Behavior Involving Minors Scale Self-Efficacy was assessed using the Aachen Self-efficacy
The Sexual Behavior Involving Minors Scale [SBIMS (38)] is a Questionnaire (31). Sample (a): Our expectation that
questionnaire measuring the frequency of specified sexual participants’ self-reported general self-efficacy would increase
behaviors during the last 6 months. To the best of our over time was not confirmed. In a sample of six participants,
knowledge, normative data are not available. Two items changes from pre-intervention to post-intervention and 1-year
concerning the frequency of sexual abuse of minors and the follow-up (Mpre = 3.5; SDpre = 0.4; Mpost = 3.7; SDpost = 0.6;
consumption of sexual exploitation material depicting minors Mfollow-up = 3.6; SDfollow-up = 0.6) were not significant (c2(2) =
were used for the purpose of this paper. Both items had to be 1.000, p = .607). Sample (b): Pre- and post-intervention data were
rated on a five-point Likert scale ranging from “never” to “daily”. further compared in a sample of 19 participants (Mpre = 3.6;
Higher scores represent more deviant sexual behavior. SDpre = 0.4; Mpost = 3.8; SDpost = 0.5). Also here, participants’
self-reported general self-efficacy did not change significantly
Sexual Fantasies and Behaviors Questionnaire (Z = −1.289, p = .197).
Since the SBIMS does not differentiate between sexual offenses
against children and those against adolescents, it was replaced by Offense-Supportive Attitudes
the Sexual Fantasies and Behaviors Questionnaire [SPV (26)]. Changes with regard to offense-supportive attitudes were
The SPV is a self-developed unpublished inventory measuring assessed using the Bumby Molest Scale (32). Sample (a): On a
the frequency of self-reported sexual fantasies of children and descriptive level, a reduction of offense-supportive attitudes was
adolescents, sexual and non-sexual contacts with minors, and the observed from pre-intervention to post-intervention to 1-year
consumption of child and youth sexual exploitation material follow-up (Mpre = 85.4, SDpre = 19.2; Mpost = 49.1, SDpost = 5.6;
during the six months prior to testing. For the purpose of this Mfollow-up = 44.3, SDfollow-up = 9.3; see Figure 1). Results of a
paper, four items were used [frequency of sexual abuse of (i) Friedman test with N = 8 show that this reduction is significant
children, and (ii) adolescents, and frequency of sexual (c2(2) = 13.067, p = .001). Bonferroni-corrected post-hoc-tests
exploitation material use of (iii) children and (iv) adolescents]. indicated a significant reduction from pre- to post-measurement
All items were rated on a six-point Likert scale ranging from 1 (p = .018; r = .49) and from pre- to follow-up-measurement (p =
(“never”) to 6 (“daily”), an example being “During the last 6 .003; r = .57) but not from post- to follow-up-measurement (p =
months, I have used child sexual exploitation material for sexual 1.0). Sample (b): The reduction of self-reported offense-
gratification”). Psychometric properties have not been assessed supported attitudes from pre- to post-intervention remained
and normative data is not available. significant in a larger sample of N = 23 with a large effect size
(Mpre = 71.0; SDpre = 19.6; Mpost = 47.1; SDpost = 7.5; Z = −3.817,
Individual Therapy Process p < .001; r = .56; see Figure 1). Median offense-supportive
Questionnaire for General and Differential Single attitude score was 69 at pre-treatment and 48 at post-treatment.
Therapy Sessions for Patients
To assess patients’ experiences with the therapeutic sessions, the
Personal Well-Being
The Symptom Checklist-90-Revised (34) was used in order to
Questionnaire for General and Differential Single Therapy
assess participants’ self-perceived overall psychological distress
Sessions for Patients [STEPP (39)] was administered. The
and results will be provided in the following. Sample (a): Changes
instrument includes three subscales (motivational clarification,
with regard to participants’ self-perceived overall psychological
problem activation, and therapeutic relationship), and comprises
distress were not significant between the three times of
of 12 items. Statements such as “What I learned today will help
measurement (N = 9; Mpre = 0.72; SDpre = 0.65; Mpost = 0.36;
me deal with my difficulties in the future” have to be rated on a
SDpost = 0.44; Mfollow-up = 0.41; SDfollow-up = 0.39; c2(2) = 3.765.
seven-point Likert-scale, ranging from 1 (“not true at all”) to 7
p = .053), but a trend was evident (see Figure 2). Sample (b):
(“absolutely true”). Higher scores on the subscales reflect greater
Results of a pre/post-comparison with N = 24 further indicate a
subjectively experienced progresses in the different domains.
statistically significant change with a medium effect size (Mpre =
Internal consistencies have been shown to be good (rtt = .76 to
0.59; SDpre = 0.49; Mpost = 0.4; SDpost = 0.39; Z = −2.159, p = .031;
rtt = .89) (39).
r = 0.31; see Figure 2). Median emotional distress score
decreased from 0.4 at pre-intervention to 0.3 at post-
Analyses intervention. The reduction remained significant when
All analyses were computed using the software SPSS version 26.0 participants receiving psychotherapy for comorbid disorders
(40). To compare mean differences between pre- and post- were excluded from the analysis (N = 19; Mpre = 0.5; SDpre =
intervention (T1 and T2) and between pre-, and post- 0.39; Mpost = 0.28; SDpost = 0.29; Z = −2.675, p = .007; r = 0.35).
intervention, and 1-year follow-up (T1, T2, and T3), Wilcoxon The Life Satisfaction Questionnaire (36) was employed to
signed-rank tests for matched pairs as well as Friedman repeated measure participants’ self-reported life satisfaction. Sample (a):
measures tests were performed. Participants’ self-reported life satisfaction did not increase

Frontiers in Psychiatry | www.frontiersin.org 5 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

FIGURE 1 | Mean offense-supportive attitude scores as measured by the BMS significantly decreased from T1 to T2 and from T1 to T3 in sample (a) and from T1
to T2 in sample (b). Error bars indicate standard errors. T1 = pre-intervention; T2 = post-intervention; T3 = 1-year follow-up.

FIGURE 2 | Participants’ mean self-perceived psychological distress score as measured by the SCL-90-R only decreased significantly from T1 to T2 in sample (b),
while no change occurred in sample (a). Error bars indicate standard errors. T1 = pre-intervention; T2 = post-intervention; T3 = 1-year follow-up.

significantly from pre-treatment to post-treatment to 1-year a statistically significant difference between the three times of
follow-up in a small sample of five participants (Mpre = 201.0; measurement (Mpre = 81.3; SDpre = 24.4; Mpost = 70.0; SDpost =
SDpre = 31.2; Mpost = 229.0; SDpost = 45.0; Mfollow-up = 238.4; 12.0; Mfollow-up = 75.6; SDfollow-up = 20.0; c2(2) = 2.242, p = .326;
SDfollow-up = 39.9; c2(2) = 4.778, p = .092). Sample (b): Also in a see Figure 3). Sample (b): In a sample of N = 25, a decrease of the
sample of 20 participants, no significant improvement regarding participants’ subjective risk perception of committing sexual
participants’ self-reported life satisfaction from pre- to post- offenses can be observed. Results further demonstrate that this
intervention was evident (Mpre = 241.5; SDpre = 45.6; Mpost = decrease was significant with a medium effect size (Mpre = 79.0;
254.9; SDpost = 38.3; Z = −1.456, p = .145). SDpre = 22.9; Mpost = 66.1; SDpost = 10.8; Z = −2.937, p = .003; r =
.42; see Figure 3). Median risk perception score decreased from
Subjective Sexual Self-Regulation 68 at pre-intervention to 61 at post-intervention.
Participants’ subjective risk perception was assessed by means of Changes regarding participants’ self-perceived ability to
the High Risk Situation Test (37). Sample (a): Concerning control deviant sexual impulses were measured using the
participants’ subjective risk perception, pre-, post-, and 1-year Coping Self-Efficacy Scale Related to Minors – Coping (38).
follow-up data were available for a small sample of N = 9. Sample (a): The increase of the participants’ self-perceived ability
However, contrary to our expectations, results do not indicate to control deviant sexual impulses permanently from pre- to

Frontiers in Psychiatry | www.frontiersin.org 6 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

FIGURE 3 | Participants’ mean subjective risk perception score as measured by the High Risk Situation Test (HRST) only decreased from T1 to T2 in sample (B),
while no change occurred in sample (a). Error bars indicate standard errors. T1 = pre-intervention; T2 = post-intervention; T3 = 1-year follow-up.

post-intervention and 1-year follow-up was not significant in a available for three participants, we refrained from conducting
sample of nine participants (Mpre = 61.2; SDpre = 12.7; Mpost = analyses with these data. Sample (b): Before the SPV (26) had
69.2; SDpost = 10.1; Mfollow-up = 67.8; SDfollow-up = 13.4; c2(2) = been put into use, the SBIMS was administered to eight
3.765, p = .152). Sample (b): Descriptively, an increase was participants to assess the frequency of (1) self-reported child or
observed in a sample of N = 23 (Mpre = 65.4; SDpre = 11.6; adolescent sexual abuse, and (2) child or adolescent sexual
Mpost = 71.3; SDpost = 9.1). This increase was, however, not exploitation material use during the 6 months prior to the
statistically significant (Z = −1.845, p = .065). Nevertheless, there assessment. Only few participants had filled in the SBIMS
was a trend in the expected direction with an increase of the halfway through the intervention (N = 6). As a consequence,
median score from 68 at pre-treatment to 75 at post-treatment. this measurement point was excluded from the analyses.
Notwithstanding this, Figure 4 depicts data from all four
Self-Reported Sexual Offenses Against measurement points. Changes over time were insignificant with
Children and Adolescents regard to both the frequency of sexual abuse offenses (Mpre = 1.4;
Results concerning deviant sexual behavior as measured by the SDpre =1.1; Mpost = 1; SDpost = 0; Z = −1, p = .317), and child and/
Sexual Behavior Involving Minors Scale [SBIMS (38)] are or adolescent sexual exploitation material use (Mpre = 2.1; SDpre =
described below. Sample (a): As follow-up data was only 1.6; Mpost = 1.3; SDpost = 0.5; Z = −1.289, p = .197).

FIGURE 4 | Neither the frequency of child/adolescent sexual abuse, nor the frequency of child/adolescent sexual exploitation material use changed significantly from
T1 to T2 as indicated by Sexual Behavior Involving Minors Scale (SBIMS) scores. Frequency was indicated on a five-point Likert Scale (1 = never; 2 = few times a
month; 3 = monthly; 4 = weekly; 5 = daily). Empty squares depict data points excluded from the analyses due to low case numbers (Ti: N = 6, T3: N = 3). Error bars
indicate standard errors. T1 = pre-intervention, Ti = intervention, T2 = post-intervention, T3 = 1-year follow-up.

Frontiers in Psychiatry | www.frontiersin.org 7 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

Results regarding deviant sexual behavior as measured by the to CSA (Mpre = 1.1; SDpre = 0.3; Mpost = 1.0; SDpost = 0; Z =
Sexual Fantasies and Behaviors Questionnaire [SPV (26)] are −1.000, p = .317; see Figure 5) nor with regard to adolescent
provided in the following. Sample (a): As follow-up data was only sexual abuse (Mpre = 1.2; SDpre = 0.6; Mpost = 1.0; SDpost = 0; Z =
available for three participants, we refrained from conducting −1.000, p = .317; see Figure 5). This result, however, does not
analyses with these data. Sample (b): The SPV was administered come as a surprise given that two participants had conducted
to assess the frequency of (i) self-reported child and adolescent contact offenses during the 6 months prior to T1 (CSAr: n = 1;
sexual abuse, and (ii) child and adolescent sexual exploitation adolescent sexual abuser: n = 1).
material use during the 6 months prior to testing. None of the (ii) Ten participants, two CSAs and eight CSEMOs had
participants filling in the questionnaire had also filled in the reported their frequency of CSEM use during the 6 months
SBIMS, meaning that in total, data on self-reported recidivism prior to the beginning of the intervention and after the
was provided by 19 subjects. Only data from participants who intervention. During the 6 months before T1, six participants
filled in the questionnaire at both pre- and post-intervention had consumed such materials, three less than once a month,
were considered for this paper. As only part of these participants and another three multiple times a week. No (further) offenses
filled in the questionnaire halfway through the intervention were reported at post-intervention or at the 1-year follow-up.
(N = 9), this measurement point was again excluded from the Results of a Wilcoxon signed-rank test indicate that the
analyses. However, to give a better picture of relapses with reduction from T1 to T2 is significant with a large effect size
regard to offline and online offenses during treatment and on (Mpre = 2.5; SDpre = 1.8; Mpost = 1.0; SDpost = 0; Z = −2.251, p =
the long term, Figures 4 and 5 depict data from all four .024; r = .50; see Figure 6). Median consumption score
measurement points (T1 = pre-intervention; Ti = intervention; decreased from 2 (less than once a month) at pre-
T2 = post-intervention; T3 = 1-year follow-up). intervention to 1 (never) at post-intervention. Eleven
(i) The SPV was administered to assess the frequency of self- participants, two CSAs and nine CSEMOs had additionally
reported CSA during the 6 months prior to testing. In total, reported their frequency of youth sexual exploitation material
data of 11 subjects were assessed before and after the use at T1 and T2. At pre-intervention, six subjects had
intervention, of three of these subjects 1-year follow-up data consumed such images and videos: one less than once a
are also available. Nine participants had consumed child or month, two one to three times a month, one once a week,
youth sexual exploitation material and the remaining two one multiple times a week, and another one daily. At post-
participants had conducted sexual offenses against children intervention, only two participants had relapsed, one of them
(less than once a month; n = 1) and adolescents (one to three less than once a month and the other one to three times a
times a month; n = 1). After the first half of treatment, at post- month. This decrease was not significant as calculated by a
intervention and at 1-year follow-up, none of the participants Wilcoxon signed-rank test (Mpre = 2.5; SDpre = 1.8; Mpost = 1.3;
reported any child or adolescent sexual offenses in the prior 6 SDpost = 0.6; Z = −1.876, p = .061), but a trend in the expected
months, suggesting that both CSAs and CSEMOs did not direction could be observed (see Figure 6). At 1-year follow-
conduct any (further) child sexual offenses. The change from up, none of the three participants had conducte d
T1 to T2 did neither reach the level of significance with regard (further) offenses.

FIGURE 6 | Participants’ frequency of self-reported child sexual exploitation

material consumption decreased significantly from T1 to T2 as indicated by
FIGURE 5 | Self-reported child and adolescent sexual abuse did not SPV scores. Changes in the frequency of self-reported adolescent sexual
decrease significantly from T1 to T2 as indicated by Sexual Fantasies and exploitation material consumption were not significant, but a trend was
Behaviors Questionnaire (SPV) scores. Frequency was indicated on a six- evident. Frequency was indicated on a six-point Likert Scale (1 = never; 2 =
point Likert Scale (1 = never; 2 = less than once a month; 3 = one to three less than once a month; 3 = one to three times a month; 4 = once a week;
times a month; 4 = once a week; 5 = multiple times a week, 6 = daily). Empty 5 = multiple times a week, 6 = daily). Empty squares depict data points
squares indicate data points excluded from the analysis due to low case excluded from the analysis due to low case numbers (Ti: N = 9, T3: N = 3).
numbers (Ti: N = 9, T3: N = 3). Error bars indicate standard errors. T1 = pre- Error bars indicate standard errors. T1 = pre-intervention, Ti = intervention,
intervention, Ti = intervention, T2 = post-intervention, T3 = 1-year follow-up. T2 = post-intervention, T3 = 1-year follow-up.

Frontiers in Psychiatry | www.frontiersin.org 8 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

TABLE 1 | Participants’ recidivism rates during treatment participation and at one-year follow-up.

Number of patients sexually abusing children Number of patients consuming sexual Number of patients conducting both
or adolescents exploitation material offense types

during treatment during one-year during treatment during one-year during treatment during one-year
(Ti-T2) f-up (T3) (Ti-T2) f-up(T3) (Ti-T2) f-up (T3)

Child and/or adolescent 0/3 -/- 0/3 -/- 0/3 -/-

sexual abusers
Child and/or adolescent 0/12 0/4 5/12 1/4 0/12 0/4
sexual exploitation material
users
Individuals with both offense 0/4 0/2 1/4 0/2 0/4 0/2
types
All participants 0/19 0/6 6/19 1/6 0/19 0/6

Recidivism rates during an average observation period of 2.4 years were calculated using SPV and SBIMS scores (26, 40). Recidivism during treatment is based on Ti and T2 scores, while
recidivism during the one-year follow-up period is based on T3 scores. Please note that the questionnaires did not cover the whole observation period and that all information is based on
self-reports and could not be compared with criminal records. f-up = follow-up.

In total, three child or adolescent sexual abusers, 12 sexual DISCUSSION

exploitation material offenders, and four men who had
conducted both types of offenses provided data on the The aim of this study was to examine the relationship between
frequency of recidivism. Table 1 gives an overview on participation in our treatment program for (potential) CSAs and
recidivism rates regarding sexual abuse and sexual CSEMOs and a variety of psychological variables. By means of
exploitation material use based on SPV (26) and SBIMS (38) self-report measures, we assessed changes from pre- to post-
scores. For the purpose of this overview, item 1 and 2 of the intervention and 1-year follow-up. Results indicate that offense-
SPV (CSA and adolescent sexual abuse) as well as item 3 and 4 supportive attitudes, emotional distress, the use of CSEM, and
(usage of sexual exploitation material depicting children and participants’ subjective risk perception of committing (further)
usage of sexual exploitation material depicting adolescents) sexual offenses decreased significantly from pre- to post-
were merged (sexual abuse of minors and usage of sexual intervention, and in the case of offense supportive attitudes
exploitation material depicting minors). Both during also from pre-intervention to 1-year follow-up. The remaining
treatment and at 1-year follow-up, none of the 19 measures of quality of life, self-efficacy, participants’ self-
participants reported to have abused a child or adolescent. perceived ability to control sexual impulses toward children
However, six participants (31.58%) had consumed sexual and adolescents permanently, and the frequency of child and
exploitation material during treatment and one participant adolescent sexual abuse, and adolescent sexual exploitation
(16.67%) during the 1-year follow-up period. One should material use did not reach a level of statistical significance,
keep in mind, however, that the two questionnaires may have although in some instances, results indicate trends in the
measured slightly different constructs due to differences in the expected direction. In the following, the main results will be
wording of the items (e.g. “comsumption of sexual exploitation discussed, and alternative explanations will be offered. In
material depicting sexual activities with minors” (SBIMS) addition, suggestions for future research directions will be made.
versus “consumption of sexual exploitation material depicting
children for sexual gratification” (SPV). Additionally, the Self-Efficacy
questionnaires did not cover the whole treatment duration. To our knowledge, self-efficacy has yet to be identified as a dynamic
Accordingly, the relapse rates presented in Table 1 may risk factor for sexual recidivism. Nevertheless, self-efficacy has been
underestimate real recidivism rates. shown to be associated with continued abstinence among drug users
(41, 42) and smokers (43) and has additionally been related to
Drop-Outs reduced dropout rates from treatment (42). Therefore, we tried to
A substantial number of participants dropped out during treatment enhance self-efficacy to increase treatment adherence. Contrary to
participation. In total, 59 out of 122 patients discontinued study our expectations, general self-efficacy did not increase significantly
participation prematurely, 13 during the diagnostic phase, 3 before during the course of the intervention. However, a closer look at the
treatment start, 31 during treatment, and 2 after the end of treatment data reveals that none of the participants had a below average sense
but before T2. Besides, 10 participants were expelled from treatment. of self-efficacy in the beginning of treatment. Instead, 63% of patients
Table 3 in the Supplementary Material provides an overview of scored within the normal range and 37% scored above average
numbers and reasons for drop-outs and expulsions during different [based on percentile ranks provided by (31)]. Accordingly, there was
phases of the therapeutic process. Reasons for drop-outs and not much room for improvement.
expulsions were determined based on participants’ self-reports or in Even though former participants did believe in their ability to
cases where patients dropped out without providing any reason, the succeed the face of adversity, individuals with a low sense of self-
therapists’ subjective perception. Whenever more than one reason efficacy may possibly participate in the program in the future. In
was applicable, the reason considered most important was recorded. that case, it would be interesting to assess if treatment

Frontiers in Psychiatry | www.frontiersin.org 9 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

participation is associated with an increase of self-efficacy and if Personal Well-Being

changes with regard to that construct are related to abstinence Our expectation that patients’ overall psychological distress
and drop-out rates. decreases from pre- to post-measurement was supported. This
does not come as a surprise since most of our patients are known
Offense-Supportive Attitudes to the justice system. In the majority of cases, participants started
Offense-supportive attitudes are an empirically supported risk the program right after they had been confronted with a house
factor for sexual recidivism (10) and are therefore considered as search or an invitation to police interrogation. In the case of
an important treatment target. Examples for offense supportive undetected offenders, strong feelings of distress may have driven
attitudes include, but are not limited to, victim-blaming, them to seek help. According to self-reports, patients felt
misperceiving social cues as sexual, or failure to take distressed for a variety of reasons, including feelings of guilt
responsibility for one’s actions. A whole treatment module in and shame, fear of being left by their partners, losing their
our treatment program is dedicated to the change of offense- children and housing, or imprisonment, and the fear that their
supportive attitudes, and throughout the intervention, cognitive offense is made public and they may be excluded socially.
restructuring is continuously being applied. Based on Quayle et Additionally, many thought they were the only individuals
al.’s recommendations for the therapeutic work with Internet sex dealing with these kinds of problems, which was perceived
offenders (44), further emphasis is placed on cognitive as onerous.
distortions that are most evident in CSEMOs. In sample (a), During therapy, we encouraged participants to develop
offense-supportive attitudes decreased significantly from pre- methods to deal with their deviant sexual fantasies, desires,
intervention to post-intervention and 1-year follow-up. From and urges. Step by step, we tried to assist them in learning how
post-intervention to 1-year follow-up, a decrease could only be to satisfy their needs in a legal and prosocial way and in
observed on a descriptive level. However, this could be due to the elaborating strategies allowing them to better handle social
fact that the sample size for the pre/post/follow-up comparison situations, and negative emotions such as anxiety, feelings of
was rather small or because the improvement that was achieved depression, et cetera. Of course, there are alternative
during therapy remained stable over time. explanations for the decrease of subjective burden. For
Also in the larger sample (b) with 23 participants, offense- instance, as can be seen in the sample description (see
supportive attitudes decreased significantly from pre- to post- Supplementary Material: Tables 1 and 2), a substantial
intervention [Mpre = 71.0 (SD = 19.6), Mpost = 47.1 (SD = 7.5)]. amount of patients suffered from at least one psychiatric
This is in line with the results of another out-patient prevention disorder in the beginning of treatment. In many cases, we
program (45), in which the score decreased significantly from recommended to undergo a second therapy to target the(se)
70.88 (SD = 17.11) to 63.30 (SD = 16.68). However, even though disorder(s). A small amount of patients indeed sought and
the control group’s cognitive distortions in this study remained received help from another mental health care professional (n
stable over time, as indicated by a Wilcoxon signed-rank test, the = 5). As this may have positively affected overall psychological
time×group interaction did not reach a level of significance (46). distress, we repeated the analysis without these patients. The
Accordingly, due to the absence of a waiting-list control group, it results, however, remained stable. Additionally, as we did not
cannot precluded that in our study, reductions also result from include a control group, we cannot preclude that the observed
time effects. changes were caused by time effects (50). This assumption would,
While at pre-intervention, our patients descriptively scored however, at the very least be plausible, given that participants
higher than a sample of incarcerated non-familial child may have adjusted to their new living conditions 2 years
molesters, nonsexual offenders, and non-offending controls following the start of treatment and—in many cases—2 years
in the community (M = 71.0 versus M = 66.0, M = 52.3, and M after (initial) contact with law enforcement authorities.
= 51.8, respectively) (47), at post-intervention, they scored Compared to psychological distress, life satisfaction did not
lower than any of the other groups. The finding that our improve significantly during the course of the intervention.
patients had lower scores than non-offending controls in the However, a closer look at the data revealed that pre-
community is especially surprising as offense-supportive intervention life satisfaction score was normal or above average
attitudes have repeatedly been associated with recidivism for the majority of patients [75%; based on general population
risk in sexual offenders in general, and also for child norms from (36)]. Further research with a larger data set will
molesters in particular (10, 48). Accordingly, the question have to reveal if the PsM’s therapeutic concept has the potential
arises of whether the decrease of offense-supportive attitudes to increase life satisfaction for those who score below-average.
as measured by the BMS really reflects changes in attitudes, or
if the difference from pre- to post-intervention is, at least Subjective Sexual Self-Regulation
partially, caused by impression management or a growing During therapy, patients identified their individual risk
understanding of the theoretical construct. In future research, situations and worked on their self-control strategies for such
it may therefore be useful to include measures of social situations. Based on earlier results (24, 26), we predicted that
desirability such as the Social Sexual Desirability Scale of the participants’ subjective risk perception of committing sexual
Multiphasic Sex Inventory (49) to control for this offenses would decrease during the course of treatment. Closer
possible effect. examination of the data reveals that risk perception at pre-

Frontiers in Psychiatry | www.frontiersin.org 10 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

intervention already was relatively low (as indicated by a score contrary to our expectations, the frequency of self-reported child
approaching the minimum score). Nevertheless, a significant and adolescent sexual abuse did not decrease from pre- to post-
decrease was observed from pre- to post-intervention, but not to intervention. This does not, however, mean that our treatment
1-year follow-up. Interpreting this result is difficult: the question concept does not have the desired effect. During the six months
arises of whether subjective risk perception decreased because prior to treatment start, only three participants had self-
participants were more comfortable with their impulse control reportedly committed sexual offenses against minors [SBIMS
skills or because they underestimated the risk posed by certain (38): n = 1; SPV (26): n = 2]. More data needs to be collected to
situations after the end of treatment. Regardless of the limited draw better conclusions on whether or not the treatment concept
informative value of the questionnaire, we nevertheless believe is associated with self-reported recidivism. Especially long-term
that it is a useful tool to determine risky situations in the data would be of special relevance to examine if potential
beginning of treatment. This information may then be used treatment effects are stable over time. This being said, one
during the therapeutic process to develop appropriate risk should keep in mind that all information gained is based on
prevention strategies. self-reports and could not be compared with criminal records. As
As compared to subjective risk perception, participants’ self- impression management or social desirability may have affected
perceived ability to control sexual impulses toward children and the results, measures such as the Social Sexual Desirability Scale
adolescents permanently did not increase significantly during the of the Multiphasic Sex Inventory (49) or the Impression
course of therapy, a finding which is in contrast to the results Management Subscale of the Balanced Inventory of Desirable
obtained by another out-patient prevention program (45). Responding (55) should be included in the test battery to control
Notwithstanding this, a trend in the expected direction was for this possible effect.
evident, and there was also relatively little room for There is an ongoing dispute on the effectiveness of child
improvement, since participants’ self-perceived self-control sexual offender therapy for CSEMOs. The core Sex Offender
abilities had already been rather high in the beginning of the Treatment Programme (SOTP) has been found to increase
intervention [sample (a): Mpre = 61.2, sample (b): Mpre = 65.4; recidivism in Internet sex offenders (56) and the question arose
highest possible score: 80]. This being said, further analyses with whether general sex offender treatments should be adapted to the
more data will have to be conducted in order to assess if needs of this specific offender group. Indeed, CSAs and CSEMOs
subjective risk perception does indeed decrease during the differ with regard to a number of dynamic risk factors, that is,
course of treatment, while self-control abilities remain stable. If factors that have been shown to correlate with recidivism risk
this was the case, treatment may have devastating consequences, and that should be addressed by therapeutic interventions [for an
as patients may underestimate the risk posed by certain overview, see (10)]. For instance, in a study in which pedophilia
situations without being able to deal better with such was assessed by means of penile plethysmography, Seto, Cantor,
situations. Despite this, results should be interpreted with and Blanchard (57) could demonstrate that as compared to
caution as both questionnaires that were used [SESM-C (51), CSAs, a higher proportion of CSEMOs is pedophilic (35%
HRST (37)] are self-report measures and therefore reflect views versus 61%, respectively). In a systematic review, it could
of the patients rather than a structured risk assessment tool further be outlined that pedophilic interest is even more
applied by mental health care professionals. pronounced in mixed offenders (58). Moreover, CSEMOs
demonstrate less offense-supportive attitudes, emotional
Self-Reported Sexual Offenses Against congruence with children, and antisocial features as indicated
Children and Adolescents by a smaller number of prior offenses and less problems with
Baseline rates of sexual recidivism in CSAs are estimated to lie supervision (27). Elliot et al. (59) further found that CSEMOs
between 13.7% and 17.5% (14, 52), while recidivism rates in a score lower on cognitive impulsivity, a component of impulsivity
sample of 541 registered CSEMOs with an average follow-up characterized by quick cognitive decision-making. As a
time of 4.1 years for new contact offenses and CSEM offenses consequence, specialized treatment protocols specifically
added up to 4% and 7%, respectively (53). At least for CSAs, it addressing relevant dynamic risk factors for CSEMOs may
was demonstrated that relapses occur less often in treated sex need to be developed (60). Based on previous results from our
offenders (13, 14, 52). While in large-scale meta-analyses, sexual research group (24, 26), we nevertheless expected that
recidivism occurred in 9.5% to 10.1% of cases (14, 52), in another participants’ frequency of self-reported child and adolescent
outpatient prevention program, 20% of CSAs relapsed based on sexual exploitation material consumption would decrease over
self-reports (45). In our sample, none of the seven CSAs or the course of the intervention. However, only the frequency of
mixed-offenders reported to have committed further child sexual child sexual exploitation material use as determined by the SPV
offenses during treatment or at 1-year follow-up (average (26) decreased significantly from pre- to post-intervention, while
observation period: 2.4 years; range: 1.04 – 4.5; SD = 0.85). the mean frequency of sexual exploitation material use depicting
Additionally, none of the twelve exclusive CSEMOs committed adolescents (SPV) or minors in generel [SBIMS (38)] remained
any first-time offenses. Levenson and Prescott (54) criticize the stable. In total, 19 subjects provided information on CSEM-
focus on absolute measures of recidivism (i.e., relapse vs. no related relapses. During treatment, six participants relapsed at
relapse) and suggest to include relative measures, such as changes least once, among them five exclusive CSEMOs and one mixed
in the frequency of relapses in treatment evaluations. However, offender. Additionally, 1 year after the end of treatment, one

Frontiers in Psychiatry | www.frontiersin.org 11 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

exclusive CSEM offender who had already relapsed during follow-up data is especially sparse, meaning that at the
treatment reported to have had a relapse (16.67%). In other moment, it is not possible yet to predict long-term changes.
words, during the average observation period of 2.4 years, Additionally, dropouts may have resulted in attrition bias. Due to
31.58% of subjects re-offended at least once. In another the long treatment length of our program (approximately 2 years
outpatient prevention program, 91% of individuals with a for the majority of patients), some of the participants dropped
history of CSEM offenses committed further online offenses out for personal reasons (e.g., move to another city, incarceration
during the one year treatment period (45). Interestingly, the or need to be transferred to an inpatient therapy setting).
self-reported recidivism rates reported in our paper and in the Moreover, some other participants dropped out for reasons
paper of this other prevention program are substantially higher that we cannot ascertain. Dropouts occurred in all phases of
than the recidivism rates in studies, in which recidivism was the treatment program, including the diagnostic phase, the
measured using criminal records. This finding is in line with the intervention, or in-between the end of the intervention and the
literature suggesting that both institutionalized sex offenders and 1-year follow-up data collection. Given that the majority of
non-incarcerated paraphiliacs disclose an enormous amount of patients dropped out due to motivational reasons, it may be
undetected sexual aggression in self-report studies (61, 62). necessary to put a stronger emphasis on behavioral techniques
Interestingly, the number of confessed sex crimes and nonsex enhancing motivation during the diagnostic phase and in the
crimes in such studies even seems to exceed the number of beginning of the intervention.
registered offenses (63). We therefore hypothesize that CSEM- Relapse rates may have been underestimated since the two
related recidivism in the literature may often be underestimated questionnaires used to assess recidivism did not cover the whole
as a result of the assessment method used. Notwithstanding this, treatment period. What is more, they may have measured
we cannot preclude that treatment may have increased slightly different constructs due to differences in the wording
recidivism in online offenders as in the case of the core SOTP of the items. Furthermore, social desirability in the form of over-
(56). Moreover, patients may have disclosed a small proportion reporting desirable or under-reporting undesirable cognitions,
of their lapses only since they may have been afraid that their feelings, or behavior may have affected the results of the study.
therapists break confidentiality in the case of frequent relapses. This tendency may have additionally been reinforced as the same
Initially, the PsM was designed as a 1-year program with three therapists performed both the treatment and the accompanying
times of measurement (T1: pre-intervention, Ti: intervention, scientific research. Unfortunately, since we do not have access to
T2: post-intervention). Accordingly, questionnaires assessing criminal records, we cannot compare the information given by
relapses during the 6 months prior to assessment were used to the patients. Moreover, one could argue that participants are not
cover the whole treatment duration. However, due to strong completely open as they may fear that the information confessed
interindividual differences regarding general cognitive ability, is passed to the police. However, a substantial amount of patients
motivational factors, and dynamic risk factors, treatment length voluntarily brought their indictments and some even confessed
had to be adapted. Additionally, a follow-up measurement one past offenses which are, according to their own testimony,
year after the end of treatment was included. Consequently, the unknown to the police and the prosecution authorities. This
results reported in this paper do not cover the whole observation suggests that participants believe in confidentiality and trust in
period. To avoid such gaps in the future, we will administer the our treatment team and is in line with the finding that
SPV half-yearly. Furthermore, we started to include a second participants judge the therapeutic relationship to be good and
relative measure of recidivism (changes in the intensity of sexual supportive [as measured by the Questionnaire for General and
offenses), since the combination of the two measures frequency Differential Single Therapy Sessions for Patients (39)].
and intensity offer a means to evaluate improvement in outcome Another limitation of the study was that we did not include a
more precisely (54). waiting-list group to control for effects of time. As a
consequence, we cannot preclude that significant changes over
time are attributed to random factors such as spontaneous
LIMITATIONS remission or regression to the mean rather than the described
treatment program (50). However, due to the patients’ self-
The empirical results reported in this paper should be considered reported high psychological strain, the potential detrimental
in the light of some limitations and must therefore be interpreted consequences of not receiving treatment, and long waiting
with caution. For instance, due to inclusion criteria imposed by times of approximately 2 years, we have decided it would be
the authors, the results reported in this paper cannot be unethical not to offer treatment to everybody in need. Changes
generalized to all CSAs and CSEMOs in forensic psychiatry. over time may have also been affected by contemporaneous
Furthermore, due to the low sample size, even external validity treatments of comorbid disorders. As can be seen in the
with regard to (potential) out-patient offenders who participate sample description, a substantial amount of patients suffered
voluntary, are intrinsically motivated, willing to change, and from comorbid disorders, especially affective and/or personality
have a high self-reported degree of psychological strain is limited. disorders. As the PsM therapy concept is not disorder-specific,
Because of changes in the data collection methodology, not all but does instead address dynamic risk factors related to sexual
participants filled in every questionnaire and not all recidivism, comorbid patients are often recommended to
questionnaires were assessed at all four points in time. The undergo an additional disorder-specific psychotherapy. Some

Frontiers in Psychiatry | www.frontiersin.org 12 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

of them were also getting treated medically, e.g. with SSRIs in issued. Written informed consent was obtained from
case of an existent affective disorder. Although this concerned all participants.
only a small number of patients, comorbid disorders as well as
medical and psychological treatment may have had confounding
effects on our results. Even though accompanying psychiatric
treatment did not seem to be a confounding variable with regard
AUTHOR CONTRIBUTIONS
to overall emotional distress, we suggest to address this question IM, TW, and JM conceived the topic of the paper. TW and IM
more deeply in future research with larger sample sizes. conducted literature searches, performed statistical analyses and
wrote the first draft of the manuscript. PF, KJ, LK, and JM
critically revised the manuscript and approved the final version.
CONCLUSION
In this paper, we examined the relationship between the
participation in our treatment program for CSAs and CSEMOs
FUNDING
and a variety of psychological variables. Over time, offense- The authors would like to thank the Lower Saxony State Office
supportive attitudes, self-reported child sexual exploitation for Social Affairs, Youth and Family, the Human Medical Center
material use, emotional distress, and participants’ subjective Göttingen, and the Asklepios Psychiatric Clinic Göttingen for
risk perception of committing (further) sexual offenses reduced their financial support enabling this work. We further
significantly. During an average observation period of 2.4 years, acknowledge support by the Open Access Publication Funds of
six out of 19 online offenders relapsed, while no further offline Göttingen University and are grateful for our patients who
offenses were reported. Although the results provide first consented that their data is being used for research purposes.
evidence for a relationship between treatment participation,
self-reported recidivism and psychological well-being, results
remain preliminary and must be interpreted with caution.
Sample sizes were small, no waiting-list control group was ACKNOWLEDGMENTS
included and participants were a subgroup of sex offenders
The authors would like to thank the Lower Saxony State Office for
with specific characteristics. Further research with a larger
Social Affairs, Youth and Family, the Human Medical Center
sample and a different research design will be necessary before
Göttingen, and the Asklepios Psychiatric Clinic Göttingen for
firm conclusions can be drawn.
their financial support enabling this work. We further
acknowledge support by the Open Access Publication Funds of
Göttingen University and are grateful for our patients who
DATA AVAILABILITY STATEMENT consented that their data is being used for research purposes.
Many thanks also to our diligent interns for their invaluable
The datasets generated for this study will not be made publicly assistance during the preparation of this paper and to Edy
available. Our data are highly sensitive (deviant sexual interest/ Beteran and David A. Martin for their linguistic support.
child sexual offenses) and cannot be anonymized.

SUPPLEMENTARY MATERIAL
ETHICS STATEMENT
The Supplementary Material for this article can be found online
This study was reviewed by the ethics committee of the at: https://www.frontiersin.org/articles/10.3389/fpsyt.2020.
University Medical Center Göttingen and a positive vote was 00088/full#supplementary-material

REFERENCES 4. Wolak J, Finkelhor D, Mitchell KJ. Child pornography possessors: trends in

offender and case characteristics. Sex Abus (2011) 23:22–42. doi: 10.1177/
1. World Health Organization. Report of the Consultation on Child Abuse 1079063210372143
Prevention. Geneva: WHO (1999). Available at: http://apps.who.int/iris/ 5. Wolak J, Finkelhor D, Mitchell KJ. Trends in Arrests for Child Pornography
handle/10665/65900. Possession: The Third National Juvenile Online Victimization Study (NJOV-
2. Great Britain Department of Education. Working Together to Safeguard 3). In: Durham, NH Crimes Against Child Res Cent (2012). p. 1–6. Available
Children: A Guide to Inter-agency Working to Safeguard and Promote the at: https://scholars.unh.edu/ccrc/46/.
Welfare of Children. London: Great Britain Department of Education (2015). 6. Cooper A. Sexuality and the internet: surfing into the new
Available at: https://assets.publishing.service.gov.uk/government/uploads/ millennium. CyberPsychol Behav (1998) 1:187–93. doi: 10.1089/cpb.
system/uploads/attachment_data/file/779401/Working_Together_to_ 1998.1.187
Safeguard-Children.pdf. 7. Alanko K, Salo B, Mokros A, Santtila P. Evidence for heritability of adult
3. Webster S, Davidson J, Bifulco A, Gottschalk P, Caretti V, Pham T, et al. men’s sexual interest in youth under age 16 from a population-based
European online grooming project - final report. Brussels: European extended twin design. J Sex Med (2013) 10:1090–9. doi: 10.1111/
Commission safer internet plus programme. (2012). jsm.12067

Frontiers in Psychiatry | www.frontiersin.org 13 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

8. Dombert B, Schmidt AF, Banse R, Briken P, Hoyer J, Neutze J, et al. How offline sex offenders against children. Arch Sex Behav (2015) 44:45–66. doi:
common is men’s self-reported sexual interest in prepubescent children? J Sex 10.1007/s10508-014-0270-x
Res (2016) 53:214–23. doi: 10.1080/00224499.2015.1020108 28. Schulz T, Hofter C, Müller JL. Prä vention sexuellen Missbrauchs
9. Neutze J, Osterheider M. (2015). MIKADO - Missbrauch von Kindern: Therapiemanual zur Arbeit mit (potentiellen) Tä tern. Beltz (2017).
Aetiologie, Dunkelfeld, Opfer - ZentraleErgebnisse des Forschungsverbundes. Available at: https://www.beltz.de/fachmedien/psychologie/buecher/
Regensburg: Abteilung für Forensische Psychiatrie und Psychotherapie, produkt_produktdetails/33414-praevention_sexuellen_missbrauchs.html.
Universität, Available at: http://www.mikado-studie.de/tl_files/mikado/ 29. Aster M, Neubauer A, Horn R. Wechsler Intelligenztest für Erwachsene WIE.
upload/MiKADO_Zusammenfassung.pdf. Deutschsprachige Bearbeitung und Adaption des WAIS-III von David
10. Mann RE, Hanson RK, Thornton D. Assessing risk for sexual recidivism: Wechsler. Scientific Reports. Frankfurt/Main, Ger Harcourt Test Serv (2006).
some proposals on the nature of psychologically meaningful risk factors. Sex 6:18573. doi: 10.1038/srep18573
Abus J Res Treat (2010) 22:191–217. doi: 10.1177/1079063210366039 30. Petermann F ed. (2012). WAIS-IV. Wechsler Adult Intelligence Scale, 4th ed.
11. Dilling H, Freyberger HJ. Taschenführer zur ICD-10-Klassifikation psychischer Frankfurt: M, Pearson
Störungen. Bern: Hogrefe Verlag (2016). Available at: http://www. 31. Wälte D, Ludwig-Becker F, Brandenburg U, Saß H, Petzold E, Kröger F.
psychosoziale-gesundheit.net/bb/pdf/BB_Taschenfuehrer_zur_ICD-10- Veränderung der kognitiven Selbstregulation bei Patienten mit
Klassifikation_psych._Stoerun.pdf. psychosomatischen Störungen im Verlauf von stationärer Psychotherapie -
12. Seto MC. Pedophilia and sexual offending against children: theory, assessment, eine kontrollierte Studie. In: Bassler M, editor. Leitlinien zur stationären
and intervention. Washington, DC: American Psychological Association Psychotherapie - Pro und Contra. Mainzer Werkstatt über empirische
(2008). doi: 10.1037/11639-000 Forschung von stationärer Psychotherapie. Gießen: Psychosozial-Verlag
13. Lösel F, Schmucker M. The effectiveness of treatment for sexual offenders: a (1999). p. 335–62.
comprehensive meta-analysis. J Exp Criminol (2005) 1:117–46. doi: 10.1007/ 32. Bumby KM. Assessing the cognitive distortions of child molesters and rapists:
s11292-004-6466-7 development and validation of the molest and rape scales. Sex Abus A J Res
14. Schmucker M, Lösel F. The effects of sexual offender treatment on recidivism: Treat (2008) 8:37–54. doi: 10.1177/107906329600800105
an international meta-analysis of sound quality evaluations. J Exp Criminol 33. Feelgood S, Schaefer G, Hoyer J. KV-M - Skala zur Erfassung kognitiver
(2015) 11:597–630. doi: 10.1007/s11292-015-9241-z Verzerrungen bei Missbrauchern. Trier: Leibniz Institute for Psychology
15. American Psychiatric Association. Diagnostic and statistical manual of mental Information (2008). doi: 10.23668/psycharchives.395
disorders (DSM-5®). Washington DC: American Psychiatric Association 34. Derogatis L. (1994). The SCL-90-R Symptom Checklist 90-R administration,
(2013). Available at: https://books.google.de/books?hl=de&lr=&id=- scoring, and procedures manual.
JivBAAAQBAJ&oi=fnd&pg=PT18&dq=Diagnostic+and+statistical+manual 35. Franke GH. SCL-90-R G Die Symptom-Checkliste von Derogatis–Deutsche
+of+mental+disorders+(DSM-5 ® )&ots=ceTN06OKs8&sig=dNGUCtdc Version. Göttingen Beltz Test (1995).
EXppUNwh47PLPxqGaKo. 36. Fahrenberg J, Myrtek M, Schumacher J, Brähler E. Fragebogen zur
16. Remschmidt H. The emotional and neurological consequences of abuse. Dtsch Lebenszufriedenheit (FLZ): Handanweisung, Fragebogen, Auswertungsbogen.
Arztebl Int (2011) 108:285–6. doi: 10.3238/arztebl.2011.0285 Göttingen: Hogrefe. (2000). Available at: https://scholar.google.de/scholar?
17. Putnam FW. Ten-year research update review: child sexual abuse. J Am Acad hl=de&as_sdt=0%2C5&q=Fragebogen+zur+Lebenszufriedenheit+%28FLZ%
Child Adolesc Psychiatry (2003) 42:269–78. doi: 10.1097/ 29%3A+Handanweisung%2C+Fragebogen%2C+Auswertungsbogen&btnG.
01.CHI.0000037029.04952.72 37. Marques JK, Day DM, Nelson C, Miner M, West MA. The sex offender
18. Widom CS, Czaja SJ, Dutton MA. Childhood victimizatin and lifetime treatment and evaluation project: fourth report to the legislature in response to
revictimization. Child Abuse Negl (2008) 32:785–96. doi: 10.1016/ PC 1365. Sacramento: California State Department of Mental Health (1991).
j.chiabu.2007.12.006 38. Neutze J, Seto MC, Schaefer GA, Mundt IA, Beier KM. Predictors of child
19. Nelson EC, Heath AC, Madden PA, Cooper ML, Dinwiddie SH, Bucholz KK, pornography offenses and child sexual abuse in a community sample of
et al. Association between self-reported childhood sexual abuse and adverse pedophiles and hebephiles. Sex Abus J Res Treat (2011) 23:212–42. doi:
psychosocial outcomes: results from a twin study. Arch Gen Psychiatry (2002) 10.1177/1079063210382043
59:139–45. doi: 10.1001/archpsyc.59.2.139 39. Krampen G. STEP: Stundenbogen für die allgemeine und differentielle
20. Gewirtz-Meydan A, Walsh W, Wolak J, Finkelhor D. The complex experience Einzelpsychotherapie. Göttingen: Hogrefe. (2002). Available at: https://
of child pornography survivors. Child Abus Negl (2018) 80:238–48. doi: scholar.google.de/scholar?hl=de&as_sdt=0%2C5&q=STEP+-+Stundenbogen
10.1016/j.chiabu.2018.03.031 +für+die+Allgemeine+und+Differentielle&btnG.
21. von Weiler J, Haardt-Becker A, Schulte S. Care and treatment of child victims 40. IBM Corp. (2019). IBM SPSS Statistics for Windows, Version 26.0. Armonk,
of child pornographic exploitation (CPE) in Germany. J Sex Aggress (2010) NY: IBM Corp.
16:211–22. doi: 10.1080/13552601003759990 41. Nikmanesh Z, Baluchi MH, Pirasteh Motlagh AA. The role of self-efficacy
22. Stiels-Glenn M. The availability of outpatient psychotherapy for paedophiles beliefs and social support on prediction of addiction relapse. Int J High Risk
in Germany. R Psychiatr (2010) 28:74–80. Behav Addict (2016) 6(1):e21209. doi: 10.5812/ijhrba.21209
23. Wild TSN, Fromberger P, Jordan K, Müller I, Müller JL. Web-based health 42. Elfeddali I, Bolman C, Candel MJJM, Wiers RW, de Vries H. Preventing
services in forensic psychiatry: a review of the use of the internet in the smoking relapse via Web-based computer-tailored feedback: a randomized
treatment of child sexual abusers and child sexual exploitation material controlled trial. J Med Internet Res (2012) 14:e109. doi: 10.2196/jmir.2057
offenders. Front Psychiatry (2019) 10:1–13. doi: 10.3389/fpsyt.2018.00763 43. de Hoog N, Bolman C, Berndt N, Kers E, Mudde A, de Vries H, et al. Smoking
24. Stolpmann G, Kahnt N, Müller I, Palmer S, Wernicke M, Wulf V, et al. cessation in cardiac patients: the influence of action plans, coping plans and
[Diagnostics and treatment of sex offenders: in psychiatric forensic self-efficacy on quitting smoking. Health Educ Res (2016) 31:350–62. doi:
committment according to section sign 63 StGB and in outpatient 10.1093/her/cyv100
prevention]. Nervenarzt (2017) 88:472–9. doi: 10.1007/s00115-017- 44. Quayle E, Erooga M, Wright L, Taylor M, Harbinson D. Only pictures?:
0313-3 Therapeutic work with internet sex offenders. Dorset: RHP (2006).
25. Beier KM. Pädophilie, Hebephilie und sexueller Kindesmissbrauch. Beier KM, 45. Beier K, Grundmann D, Kuhle L, Scherner G, Konrad A, Amelung T. The
editor. Berlin, Heidelberg: Springe-Verlag (2018). doi: 10.1007/978-3-662- German Dunkelfeld Project: a pilot study to prevent child sexual abuse and the
56594-0 use of child abusive images. J Sex Med (2015) 12:529–42. doi: 10.1111/
26. Schulz T, Palmer S, Stolpmann G, Wernicke M, Müller JL. Presenting a jsm.12785
treatment concept for people with a self-reported sexual interest in children in 46. Mokros A, Banse R. The “Dunkelfeld” Project for self-identified pedophiles: a
an outpatient setting. Open J Psychiatry (2016) 07:1–17. doi: 10.4236/ reappraisal of its effectiveness. J Sex Med (2019) 16:609–13. doi: 10.1016/
ojpsych.2017.71001 j.jsxm.2019.02.009
27. Babchishin KM, Hanson RK, VanZuylen H. Online child pornography 47. Marshall WL, Marshall LE, Sachdev S, Kruger RL. Distorted attitudes and
offenders are different: a meta-analysis of the characteristics of online and perceptions, and their relationship with self-esteem and coping in child

Frontiers in Psychiatry | www.frontiersin.org 14 March 2020 | Volume 11 | Article 88

Wild et al. Prevention of Sexual Child Abuse

molesters. Sex Abus J Res Treat (2003) 15:171–81. doi: 10.1023/ 57. Seto MC, Cantor JM, Blanchard R. Child pornography offenses are a valid
A:1023547316429 diagnostic indicator of pedophilia. J Abnorm Psychol (2006) 115:610–5. doi:
48. Helmus L, Hanson RK, Babchishin KM, Mann RE. Attitudes supportive of 10.1037/0021-843X.115.3.610
sexual offending predict recidivism: a meta-analysis. Trauma Violence Abus 58. Hanson RK, Bourgon G, Helmus L, Hodgson S. The principles of effective
(2013) 14:34–53. doi: 10.1177/1524838012462244 correctional treatment also apply to sexual offenders: a meta-analysis. Crim
49. Nichols HR, Molinder I. Multiphasic sex inventory: a test to assess the Justice Behav (2009) 36:865–91. doi: 10.1177/0093854809338545
psychosexual characteristics of the sexual offender. Tacoma, WA: Nichols & 59. Elliott IA, Beech AR, Mandeville-Norden R, Hayes E. Psychological profiles of
Molinder Assessments, Inc. (1984). internet sexual offenders: comparisons with contact sexual offenders. Sex Abus
50. Barnett AG, van der Pols JC, Dobson AJ. Regression to the mean: what it is J Res Treat (2009) 21:76–92. doi: 10.1177/1079063208326929
and how to deal with it. Int J Epidemiol (2005) 34:215–20. doi: 10.1093/ije/ 60. Magaletta PR, Faust E, Bickart W, McLearen AM. Exploring clinical and
dyh299 personality characteristics of adult male internet-only child pornography
51. Neutze J, Seto MC, Schaefer GA, Mundt IA, Beier KM. Predictors of child offenders. Int J Offender Ther Comp Criminol (2014) 58:137–53. doi:
pornography offenses and child sexual abuse in a community sample of 10.1177/0306624X12465271
pedophiles and hebephiles. Sex Abus A J Res Treat (2011) 23:212–42. doi: 61. Weinrott MR, Saylor M. Self-report of crimes committed by sex offenders.
10.1177/1079063210382043 J Interpers Violence (1991). 6(3):286–300. doi: 10.1177/088626091006003002
52. Gannon TA, Olver ME, Mallion JS, James M. Does specialized psychological 62. Abel GG, Becker JV, Mittelman M, Cunningham-Rathner J, Rouleau JL,
treatment for offending reduce recidivism? a meta-analysis examining staff Murphy WD. Self-reported sex crimes of nonincarcerated paraphiliacs.
and program variables as predictors of treatment effectiveness. Clin Psychol J Interpers Violence (1987). doi: 10.1177/088626087002001001
Rev (2019) 73:101752. doi: 10.1016/j.cpr.2019.101752 63. Scurich N, John RS. The dark figure of sexual recidivism. Behav Sci Law
53. Eke AW, Seto MC, Williams J. Examining the criminal history and future (2019). doi: 10.1002/bsl.2400
offending of child pornography offenders: an extended prospective follow-
up study. Law Hum Behav (2011) 35:466–78. doi: 10.1007/s10979-010- Conflict of Interest: The authors declare that the research was conducted in the
9252-2 absence of any commercial or financial relationships that could be construed as a
54. Levenson J, Prescott DS. Déjà vu: from Furby to Långström and the evaluation potential conflict of interest.
of sex offender treatment effectiveness. J Sex Aggress (2014). 20(3):257–66. doi:
10.1080/13552600.2013.867078 Copyright © 2020 Wild, Müller, Fromberger, Jordan, Klein and Müller. This is an
55. Paulhus DL. Measurement and Control of Response Bias. Measures of open-access article distributed under the terms of the Creative Commons Attribution
Personality and Social Psychological Attitudes. New York: Academic Press License (CC BY). The use, distribution or reproduction in other forums is permitted,
(1991) 1:17–59. doi: 10.1016/b978-0-12-590241-0.50006-x provided the original author(s) and the copyright owner(s) are credited and that the
56. Mews A, Di Bella L, Purver M, Ministry of Justice. Impact evaluation of the original publication in this journal is cited, in accordance with accepted academic
prison-based core sex offender treatment programme. London, England: practice. No use, distribution or reproduction is permitted which does not comply with
Ministry of Justice Analytical Series (2017). these terms.

Frontiers in Psychiatry | www.frontiersin.org 15 March 2020 | Volume 11 | Article 88