Is Preventive Treatment for Individuals With Sexual Interest in Children Viable in a Discretionary Reporting Context?

Abstract

This study explored the viability of preventive treatment services for individuals with sexual interest in children, in jurisdictions without mandatory reporting but where risk-related disclosures to authorities are permitted at therapists’ discretion. Health professionals (N = 112) were surveyed regarding their comfort, confidence, knowledge of relevant legal provisions, and personal disclosure thresholds, in relation to a hypothetical scenario of a client confiding pedophilic interest to seek help. Findings were mixed regarding implications for prevention service viability. Despite the complexities of the legal and ethical context of the study setting (New Zealand), predictions regarding professionals’ uncertainty in relation to their legal and ethical duties, and displaying a bias toward disclosing information to authorities when permitted, were not fully borne out, although pervasive knowledge inaccuracies and associated training needs were revealed. Instead, general tendencies among respondents were toward comfort, confidence, and the inclination toward maintaining client confidentiality. Yet, widespread variance within the sample, and individuals’ thresholds appearing rather unpredictable on the basis of demographic or professional variables, highlights likely barriers for potential clients in feeling safe enough to come forward. Given that preventive treatment viability in this context relies on self-referral, it is suggested that a purpose-designed preventive treatment service, with clear accessible confidentiality and reporting policies that are well within the law, could be the best way forward for viable preventive treatment in discretionary reporting contexts.

Keywords

prevention of child abuse child abuse sexual abuse offenders sexual assault intervention reporting/disclosure

Societal approaches to reducing the occurrence of child sexual abuse have traditionally focused on the tertiary level of prevention (Leclerc, Chiu, & Cale, 2016). This approach involves providing treatment to those known to authorities by virtue of having a relevant conviction, with the aim of reducing the likelihood that they will repeat the behavior (Becker & Reilly, 1999; Centers for Disease Control and Prevention [CDC], 2004). This is clearly an important justice system pursuit, and its effectiveness in terms of reducing recidivism has been supported in large-scale international meta-analytic research (e.g., Gannon, Olver, Mallion, & James, 2019). However, conviction figures from New Zealand (NZ), the setting for the current study, suggest that to restrict the focus solely to this level of prevention may drastically limit our ability to reduce child sexual abuse incidence and victimization rates: In NZ, in 2015, out of the 411 individuals who received a conviction for a sexual offense against a child, only 51 of them, or 12%, had a prior record of sexual offending against children. A further 16 (4%) had a prior conviction for sexually offending against an adult victim. This leaves the vast majority, 84%, who were previously unknown to authorities to pose a sexual risk (New Zealand Ministry of Justice, 2016).

To have any hope of preventing the major proportion of child sexual offenses, it therefore seems vital to expand efforts to temporally earlier, primary prevention initiatives. According to CDC, primary prevention refers to “approaches that take place before sexual violence has occurred to prevent initial perpetration or victimization” (2004, p. 3). CDC also differentiates in terms of the intended targets: those aimed at the general population are termed universal, whereas selected interventions are aimed specifically at those thought to have an elevated risk (CDC, 2004). Applying this terminology, the focus of the current article will be selected interventions in a primary prevention context; specifically, the idea of making services available at an earlier stage for people who experience sexual interest or attraction toward children, so as to prevent progression toward abusive behaviors. Such an approach is not a new idea (e.g., Chasan-Taber & Tabachnick, 1999) but remains relatively rare internationally. With tertiary programs in the justice sector often referred to as rehabilitation (e.g., Lloyd, Hanby, & Serin, 2014), the kind of primary, selected approach described here could suitably be termed prehabilitation. To the author’s knowledge, this term has not been previously applied in the sexual offending literature, but in medical settings it is used to describe a process of enhancing functional capacity, to enable an individual to withstand a stressful event such as surgery (Ditmyer, Topp, & Pifer, 2002). Prehabilitation in the context of child sexual abuse prevention can therefore be conceptualized as providing “strength training” of sorts to those who experience sexual interest in children, to enhance their capacity to cope with this in nonharmful ways.

An undesirable reality for settings where prehabilitation is not available is that for a problem for which empirically supported strategies have been developed (i.e., risk of sexual offending and the successful reduction of this through treatment), interventions utilizing these strategies are only accessible within the justice system. As a result, individuals at risk may find that they are unable to access any feasible options for help until after they have progressed to offending, undoubtedly having caused great harm. Given the extremely high human/social costs of the problem of sexual abuse (Maniglio, 2009), in addition to the breadth and extent of associated financial burdens such as investigation, prosecution, imprisonment, victim care, and productivity loss (Letourneau, Brown, Fang, Hassan, & Mercy, 2018), perhaps prehabilitation ought to be viewed as a rather obvious and cost-effective strategy to assist people at risk, without requiring harm to have occurred first.

What might be the scope of demand for such a service? The incidence of sexual interest in children under the age of 16 years based on confidential self-report was found to be 3.0% in a population-based cohort study of nearly 4,000 men in Finland (Alanko, Salo, Mokros, & Santtila, 2013). Other studies employing different definitions and methods have produced a range of estimates, many substantially higher (as reviewed by Lasher & Stinson, 2017). Whatever the true prevalence, it seems clear that within most communities will be a not insignificant number of people living with the awareness that they are sexually interested in children. Contrary to popular public perception which often casts such individuals as predatory (McCartan, 2004) and inevitably posing a danger to children (Jahnke & Hoyer, 2013), evidence suggests that many experience substantial distress as a direct consequence of their sexual interest (Schaefer et al., 2010), and furthermore, will come forward to receive preventive help when suitable services are available (Beier, Neutze, et al., 2009).

Prevention Project Dunkelfeld, in Germany, provides a working example of how prehabilitation can operate: Dunkelfeld offers publicly funded comprehensive assessment and treatment services on a self-referral basis to individuals with pedophilia living in the community; both those who have perpetrated undetected child sexual abuse, and those who have not but fear they may. Treatment is offered in group and individual formats, and encompasses cognitive-behavioral, sexological, and medicinal approaches (Beier, Ahlers, et al., 2009). Importantly, these services are provided under a guarantee of complete confidentiality (Beier et al., 2015), meaning that clients can feel safe to be open and honest to address the full range of their problematic sexual interests and related treatment needs. This situation addresses one of the most common and problematic barriers to help-seeking among those who experience sexual interest in children—the fear of potentially dire legal and social consequences if they were to confide in someone to seek help (Levenson, Willis, & Vicencio, 2017). The Dunkelfeld program can therefore be viewed as having extended (in its setting) the societal response to the problem of child sexual abuse to include typically neglected selected primary prevention. Dunkelfeld was launched in 2005, with evaluation research indicating success at measurably reducing several critical risk factors (Mann, Hanson, & Thornton, 2010) relative to a comparison group, including emotional deficits, offense-supportive cognitions, and sexual dysregulation (Beier et al., 2015).

So why is the Dunkelfeld model not followed everywhere? Key to its viability in Germany is the specific legal context. Under German law, therapeutic confidentiality is explicitly protected—it is an offense for therapists in nonforensic settings to divulge confidential information about their clients, including regarding past or planned child sexual abuse perpetration (Beier, Ahlers, et al., 2009). Although client confidentiality in health care settings is a well-established principle internationally (Beauchamp & Childress, 2013), unlike Germany, many jurisdictions have over time adopted legislation that effectively rules out the confidentiality guarantee offered at Dunkelfeld. In fact, in many places the direct opposite is in effect—in certain circumstances disclosure of client information to authorities is actually mandated. In the United States, for example, all states have identified professionals required to report known or suspected child sexual abuse to authorities, typically including a range of health and mental health professionals (Child Welfare Information Gateway, 2016). Some states including Maryland, Connecticut, and Hawaii, also mandate reporting of risk of potential abuse when specific criteria are met. Likewise, all states and territories in Australia have adopted some form of mandated reporting, and in some, such as New South Wales, this extends to circumstances of reasonable suspicion that a child is at risk (Australian Institute of Family Studies, 2017). Although, as Lasher and Stinson (2017) note, the intent of such laws has been to enhance child protection, a consequence has been to remove avenues to help-seeking for the very people who may pose a risk. In such settings it is difficult to see how face-to-face therapeutic prevention services based on self-referral like Dunkelfeld could operate.

However, in other jurisdictions in the world, the legal context in relation to therapeutic confidentiality is not so polar, with neither mandated reporting, nor as strong protection for therapeutic confidentiality. One example where the situation falls in the middle of these two extremes is NZ, the setting for the current study. The central focus of this study is whether a Dunkelfeld-like service could potentially operate in such a jurisdiction, given the legislative context, as well as the prevailing views and norms of those practicing within it. Therapeutic confidentiality under NZ law will now be examined closely, to understand the research context as well as facilitate comparison with jurisdictions with similar provisions.

In NZ, privacy of consumer health information is primarily governed by the Health Information Privacy Code 1994. Under this code, the relevant rule (Rule 11) is that health information must not be disclosed by an agency holding it, unless one of a list of exceptions applies—one of which is if disclosure is necessary to prevent or lessen a serious threat. If an exception applies, disclosure (to the extent necessary) is “permitted,” but is not mandated. Therefore, health professionals who perceive that a client poses a risk to others must exercise considerable discretion: first, in judging whether the risk is sufficient to meet the exception clause, and second, if so, whether or not to make the permitted disclosure. One rationale for opting not to disclose could be, for example, wishing to preserve the therapeutic relationship so as to reduce and/or manage the perceived risk in that way. Another relevant piece of legislation is the Oranga Tamariki Act 1989,¹ the provisions of which relate specifically to child protection, but are not specific to information collected in health contexts. Under Section 15, “Any person who believes that any child or young person has been, or is likely to be, harmed (whether physically, emotionally, or sexually), ill-treated, abused, neglected, or deprived may report the matter to [authorities].” The use of the word “may” as opposed to “must” clarifies that this provision too permits disclosures, but does not mandate them. Finally, the Crimes Act 1961 sets out liability for those with knowledge of a risk of sexual assault to a child and fail to protect them; however, this applies only to members of the same household as the victim or staff of an institution where the victim resides. As such, this mandate would typically not apply to professionals providing Dunkelfeld-style services.

The sum effect of the NZ legislative situation, therefore, is that there is not mandated reporting in relation to clients experiencing sexual interest in children. Disclosure of client information is permitted when certain criteria are met, but remains discretionary. There is one further legislative clause with direct relevance: Section 16 of the Oranga Tamariki Act 1989 states that no civil, criminal, or disciplinary proceedings shall lie against those who make a disclosure of information concerning a child, provided the information was disclosed in good faith. This means that although professionals are able to exercise discretion regarding disclosures, they are professionally protected if they opt to disclose when permitted. In contrast, no such protection is afforded for good faith decisions to preserve confidentiality.

Outside of legislation, it is also important to consider case law, and the concept of duties that may apply under the tort of negligence. In other words, whether a health professional could be liable for failing to protect a victim despite the legislative discretion, similar to the duty established in the United States by the well-known Tarasoff decision (Tarasoff v. Regents of the University of California, 1976). Such a duty has not yet been found to apply in NZ; however, a number of cases it has been tested in have outlined circumstances in which it potentially may (Harris, 2013). Also weighing into discretionary decisions are professional codes of ethics, which may hold professionals to a higher ethical standard pertaining to client confidentiality than the law requires. Yet, some also make explicit reference to the paramountcy of the protection of children, stating that this should be given precedence over other considerations (e.g., Code of Ethics for Psychologists Practicing in Aotearoa/New Zealand; New Zealand Psychologists’ Board, 2002; s 1.5.1). Finally, employment provisions may also bear relevance, for example, where internal policies have been established regarding reporting, which staff may be considered bound to adhere to despite the lack of statutory mandate. In some circumstances, therefore, professionals may have less discretion in these kinds of decisions than initial appearances.

All in all, it appears that in settings where the legal context does not involve polar provisions such as are in effect in Germany, the United States, and Australia, the situation may in fact be highly complex. Health professionals, usually without a background of legal training, may find it difficult to wade through varying provisions from multiple sources to have a clear understanding of what their duties and responsibilities actually are. Furthermore, this lack of clarity, coupled with the likely heavy-felt responsibility of having discretion over decisions pertaining to child safety (considered of inherent importance in most societies), may promote a certain risk aversion among health professionals, particularly given (in NZ) the unbalanced professional protection offered for decisions when they are in the direction of disclosure. It is finally worth highlighting an issue arising from permitted disclosures including situations of perceived risk. Although the lack of a reporting mandate may at first glance appear to be helpful to prehabilitation viability, mandated reporting provisions where they exist tend to require the higher threshold of known or suspected actual abuse. The discretion afforded NZ practitioners can therefore “cut both ways,” allowing either more or less conservative decision-making regarding disclosure relative to many mandatory reporting jurisdictions. Given this, discretionary contexts could therefore equally be viewed as detrimental to prehabilitation viability.

The focus of the current study is on exploring health care professionals’ understanding and navigation of client confidentiality in a discretionary reporting environment (NZ), with specific reference to the scenario of a client seeking help in relation to sexual interest in children. Implications for the viability of Dunkelfeld-style prehabilitation services will be considered. In addition to the exploratory aim, two specific predictions were made on the basis of the above-described legal and ethical context: health care professionals working in this context would (a) experience uncertainty regarding their legal and ethical duties when responding to a relevant client scenario and (b) display a bias toward opting to disclose client confidences to authorities when permitted to. Evidence relating to these predictions was evaluated using a survey targeted at health care professionals working in roles accessible to the public via self-referral.

Method

Design and Procedure

The current study involved an anonymous online self-report survey. To meet eligibility criteria, respondents needed to be a qualified health professional, working in NZ in a role accessible to the public via self-referral. Specifically targeted professions included general medical practitioners, psychologists, counselors, and psychotherapists. Participants were recruited through advertisements sent via a number of relevant professional groups (e.g., New Zealand College of Clinical Psychologists), and in some cases via their employing clinics (e.g., campus-based student health and/or counseling services around NZ). Interested and eligible individuals voluntarily accessed a consent form and the survey by clicking on a provided link to the Qualtrics platform. A small incentive (NZ$10 voucher) was offered to participants in appreciation of their time. This study was approved by the University of Canterbury Human Ethics Committee (Ref.: HEC2017/48/LR-PS).

Participants

A total of 112 participants completed the full questionnaire. Participants included males (19%) and females (81%) aged between 23 and 74 years (M = 46.49, SD = 10.96). The ethnic breakdown² was as follows: 67% NZ European/Pākeha, 16% other Caucasian, 9% Asian, and 5% NZ Māori, with the remaining 3% being of other ethnicities. Represented professions included predominantly psychologists (47%), general medical practitioners (24%), and counselors (19%), with the remaining 10% including small numbers of psychotherapists, psychiatrists, social workers, hospital doctors, and nurse practitioners. The level of professional experience ranged from less than one year (7%) to upward of 50 years for one individual, with an average of 15.83 years (SD = 10.68). With regard to specific relevant experience, 37% of participants reported that they had experienced, at least once, a client of their professional practice disclosing to them a sexual interest in children. For 44% of those, this had happened only once; it had happened between 2 and 3 times for 41%; with the remaining few (15% or six individuals) reporting more than three such client experiences. Eligibility criteria did not require this specific experience, as the questionnaire content (see below) relied largely on hypothetical responses to written scenarios.

Questionnaire

The questionnaire completed by participants had three main components, in addition to basic demographic and professional experience information. These were as follows: comfort and confidence with a described hypothetical professional scenario involving client disclosure of pedophilic interest, knowledge regarding legal and ethical obligations arising from the described professional scenario, and personal thresholds for breaching client confidentiality for risk management reasons given the described professional scenario. The comfort, confidence, and knowledge variables were designed to yield information relevant to the prediction of health professional uncertainty, whereas the threshold variables were to consider the question of bias. To avoid missing data, all fields were set to require a response before the respondent could advance to the next question. The entire survey took approximately 10 to 12 min to complete. Survey contents regarding each of the three major components are outlined in more detail below (available in full upon request).

Comfort and confidence

Participants were provided a brief scenario involving a hypothetical client disclosing during an appointment that they experience sexual attraction toward a child or children. It was specified that the purpose of the client making this disclosure was so as to seek help from the professional (either directly or via an appropriate referral), and that there was nothing to suggest the client had ever committed an abusive act. Participants were then asked to respond regarding how comfortable they would feel facing this scenario, and (separately) how confident they felt in their ability to appropriately deal with such a disclosure from a client, on 7-point Likert-type scales. Comfort and confidence were assessed separately to disentangle the two, as they were considered likely to be related but separate concepts, and equally of interest. Confidence was further separated into two scales, to allow different responses regarding how confident participants would feel within the session, and overall (i.e., given time to consult, etc.). This distinction was made in recognition that in normal practice, at least in the absence of acute risk, professionals would typically have time to consider their response and consult as widely as desired before making decisions regarding how to act. Participants who indicated any level of discomfort and/or low confidence (i.e., below the “neutral” rating) with the scenario were then asked to select any that applied from a list of six potential reasons (e.g., lack of experience with this kind of client/issue, dislike or moral opposition to a client with this issue); participants could also specify additional reasons that applied to them in free text form.

Knowledge

Participants’ knowledge regarding relevant legal and ethical obligations was assessed in two ways. The first was a self-assessment involving a 5-point Likert-type scale ranging from I do not know anything about this to I consider myself an expert in this, with reference to legal and ethical duties that would arise if facing the scenario. The second involved a brief quiz, consisting of four multi-choice questions addressing participants’ understanding of the following: current NZ legislation relating to the existence (or otherwise) of mandated reporting of risk, specifically applied to confidential health information³ and more generally applied to harm or risk of harm toward children,⁴ and legislation relating to immunity (or otherwise) from professional disciplinary proceedings with respect to such reporting.⁵ To avoid an oversubscribed I don’t know category, questions utilizing a true/false format included two additional categories designed to force a choice for those who were unsure while still allowing uncertainty to be expressed (i.e., Not sure, but probably true and Not sure, but probably false). For all knowledge-related questions, participants were asked to rely on their existing knowledge only despite recognition that in practice they may choose to check or consult with external sources.

Thresholds

Following the knowledge-related questions, participants were presented with brief information regarding the actual NZ legal situation: Specifically, that it is permitted but not mandated to disclose otherwise confidential health information to authorities in circumstances where there is a perceived serious threat, and that consequently, professionals who perceive such a threat must exercise discretion in deciding whether to make a report or maintain client confidentiality. It was considered that thresholds for this decision would likely vary; that is, some individuals may lean toward making a report whenever legally permitted, some in contrast may tend toward not doing so, whereas others may have varying tendencies somewhere in between these positions. In addition to evaluating the prediction of bias, individual thresholds were of interest given the pertinence of concerns with therapeutic confidentiality among potential help-seekers (Levenson et al., 2017). Like knowledge, participants’ thresholds for making a report were assessed in two ways, as described below.

The first measure of thresholds was a self-assessment, in which participants indicated on a 7-point scale what they thought their general tendency would be if faced with the client scenario. Response options ranged from definitive choices at either pole (make a risk-related disclosure, or maintain client confidentiality) even if that would mean breaking the law, to the same definitive choices unless that would break the law, to a more moderate “lean toward” one decision or another but be willing to consider the opposite action depending on the specific details of the case, to finally the scale midpoint of “No general tendency—it could go either way depending on the specific details of the case.” Participants who selected any response but the midpoint were asked to indicate their main reasons for their self-identified tendency by clicking on any that applied from a list of six and/or entering their own additional reasons using free text entry. Participants whose response indicated that their decision would depend on the specific details of the case (i.e., Points 3, 4, or 5 on the Likert-type scale) were asked to indicate what kinds of details they would likely take into account.

The second method of assessing participants’ thresholds involved dichotomous forced-choice responses (“disclose to authorities” or “maintain confidentiality”) for a series of variations on the key scenario (i.e., a client disclosing sexual interest in children as a means of seeking help), intending to depict varying degrees of harm likelihood. Twenty scenarios were created based on combinations of the presence or absence of the following details regarding the hypothetical client: whether their attraction was directed at a specific identifiable child (e.g., their boss’ child), whether they lived in a household with any children, whether there were foreseeable instances of unsupervised contact with children, and whether there was evidence of problems with self-control (e.g., substance misuse, impulsive tendencies) or conversely evidence of positive self-control. For example, a scenario at the lower end of the spectrum (with just one of the four above-described elements) included the following additional details: “No specific current target of attraction reported; Not living with children; Client reports occasional unsupervised contact with children (e.g., friends’ children); Evidence of positive self-control.” A scenario intending to suggest a higher likelihood of harm (three of the four elements) specified,

Specific self-reported target of attraction is boss’s child; client reports occasionally being left alone with this child; Living in a household with children, however denies any attraction to them; Evidence of problems with self-control (e.g., substance misuse, or impulsive tendencies).

To avoid survey fatigue, each participant responded to a randomly selected 12 of the 20 scenarios, presented in randomized order to prevent order effects. To then obtain an estimate of each participant’s threshold for disclosing, their 12 responses (either 0 for Disclose to authorities or 1 for Maintain confidentiality) were summed, to create a continuous variable with a possible range from 0 (i.e., for those who indicated they would choose to disclose for every scenario presented to them) to 12 (for those who would maintain confidentiality each time).

Planned Analyses

To fulfill the exploratory aim of the study, as well as to address the predictions of uncertainty and bias, sample distributions were analyzed on each of the key dependent variables: comfort and confidence, self-assessed knowledge of legal and ethical duties, accuracy of knowledge, self-assessed tendency regarding whether to disclose to authorities or not, and the summed estimate of disclosure threshold. Thresholds were of particular interest in considering the viability of Dunkelfeld-style prehabilitation given the discretionary reporting context; it was therefore next planned to investigate whether demographic variables (i.e., gender, ethnicity, age, experience, profession) or the other key variables (i.e., comfort, confidence, knowledge) bore any relationship with self-assessed or estimated thresholds.

Results

Health Professional Comfort and Confidence

The mean rating on the 7-point Likert-type scale regarding comfort with the hypothetical scenario (with 4 being neutral and higher scores indicting increasing comfort) was 4.51 (SD = 1.63). Mean confidence was slightly higher at 4.79 in session (SD = 1.44), and higher again for the scenario where participants could have time to consult (M = 5.38; SD = 1.26). The sample response distributions are depicted in Figure 1. As can be seen, responses were mixed but overall tended toward comfort and confidence (as opposed to discomfort and lack of confidence). More than half of the sample (59%) responded with a score above neutral for comfort (i.e., 5, 6, or 7), indicating some level of comfort between somewhat to very comfortable. The corresponding proportions for confidence were 69% in session, and 80% (i.e., four out of every five respondents) if given time to consult.

Figure 1.

Distribution of responses on 7-point scale regarding comfort and confidence with the client scenario (N = 112).

Among the n = 51 participants who reported any level of discomfort and/or lack of confidence with the scenario, the highest endorsed reason was Lack of experience with this kind of client/issue (n = 40; 78%), followed by Uncertainty regarding my legal obligations in this scenario (n = 20; 39%), and Lack of knowledge about this client issue (n = 18; 35%). Several respondents used the free text option to add a perceived lack of available local services to refer such a client where they could receive the kind of help they needed.

Health Professional Knowledge

Self-assessed knowledge

Figure 2 displays the sample response distribution on the 5-point scale for self-assessed knowledge regarding legal and ethical duties that would arise from the hypothetical scenario. As can be seen, the sample tended to rate their knowledge as at least sufficient, with three quarters selecting the midpoint of 3 (I know enough about this to get by) or higher. The modal rating was 3, and the mean was 3.03 (SD = 0.84).

Figure 2.

Distribution of responses on 5-point scale regarding self-assessed knowledge of legal and ethical duties that would arise with the client scenario (N = 112).

Accuracy of knowledge

Figure 3 displays the sample response distributions for the first two multi-choice questions, designed to assess the accuracy of participants’ knowledge relating to the existence or otherwise of mandated reporting in their professional jurisdiction (NZ). As can be seen, for both questions (asked in different ways and relating to two different pieces of legislation), the majority of participants appeared to mistakenly believe that mandated reporting laws were in effect (75% for the first question; and 67% for the second, increasing to 79% when those who were not sure but suspected so were included).

Figure 3.

Distribution of responses on knowledge accuracy questions regarding the existence of mandated reporting laws in their jurisdiction. Question 3a: “Under New Zealand law, if a client confidentially provides information to a health professional like yourself that indicates the client poses a serious threat to the life or health of another person, the health professional . . .” Question 3b: True or false? Under New Zealand law, any person who believes that any child or young person has been, or is likely to be, harmed (whether physically, emotionally, or sexually), ill-treated, abused, neglected, or deprived must report the matter to authorities.

The next two multichoice questions were designed to assess the accuracy of participants’ knowledge of provisions relating to professional liability when risk-related disclosures of confidential information are made in contexts of perceived risk. As depicted in Figure 4, knowledge accuracy was higher for these questions than for those shown in Figure 3, with around half of participants accurately understanding that they would be professionally protected if they chose to make a permitted risk-related disclosure, but would not be if they applied their discretion to maintain client confidentiality (49% and 54%, respectively). This increased to more than three quarters (76% and 84%) when those who were not sure but leaned toward the correct answers were included.

Figure 4.

Distribution of responses on knowledge accuracy questions regarding legislative provisions relating to protection from professional liability. Question 4a: True or false? Under New Zealand law, if a health professional like yourself reports to the relevant authorities their belief that a child is likely to be sexually abused, the health professional is protected from disciplinary proceedings against them with respect to this disclosure. Question 4b: True or false? Under New Zealand law, a health professional like yourself who does not report to authorities client information indicating a potential risk to another person (e.g., due to the principle of confidentiality) is protected from disciplinary proceedings against them with respect to this nondisclosure.

Health Professional Thresholds

Self-assessed thresholds

Figure 5 displays the sample response distribution on the 7-point scale regarding what participants considered their general tendency would be if faced with the above-described scenario. As can be seen, overall, the sample tended to self-rate as leaning toward maintaining client confidentiality, with 57% endorsing a response on that side of neutral (5, 6, or 7), compared with 14% who endorsed a response on the other side of neutral (3 or 2; no one selected 1) indicating a tendency toward making a risk-related disclosure. The modal rating was 5 (Lean toward maintaining client confidentiality if the law permitted it, however, would consider making a risk-related disclosure to the relevant authorities depending on the specific details of the case), and the mean was 4.55 (SD = 0.98).

Figure 5.

Distribution of responses on 7-point scale regarding self-assessed general tendency in relation to making a risk-related disclosure versus maintaining client confidentiality (N = 112).

Among the n = 16 whose response on the self-assessed threshold item indicated any level of tendency toward making a risk-related disclosure (i.e., those selecting anything that side of neutral; a rating of 3 or 2), the highest endorsed reason was Concern for the people who could be harmed if I don’t disclose (n = 14; 88%), followed by Following what I’ve been taught from training and/or supervision (n = 9; 56%), and Protecting myself professionally (n = 8; 50%). Conversely, for the n = 64 individuals who reported tending toward maintaining client confidentiality, It seems like the best way to mitigate the risks was the highest endorsed reason (n = 50; 78%), followed by Concern for the greater good if I disclose (n = 32; 50%), and Following what I’ve been taught from training and/or supervision (n = 24; 38%).

Among the n = 97 participants (87% of the sample) whose self-assessed threshold rating indicated that their response to the scenario would depend on the specific details of the case (3, 4, or 5), all nine provided options regarding the kinds of details they would likely take into account were highly endorsed, with frequencies above 90% for five of the nine, and 66% for the lowest endorsed option. This indicates that most professionals would likely take into account a wide range of factors in deciding how to respond, as opposed to a small number or any single case detail. The nine provided responses are listed below in descending order of endorsement frequency: If the client in fact had previously committed an abusive act (95%), Whether or not the client has noted a particular child or children that they experience attraction toward (93%), Whether the client lives with any children (93%), The client’s risk level (if known) (92%), Whether the client has potential access to children through settings such as work or community groups (91%), The client’s motivation to engage in treatment regarding this issue (87%), The client’s patterns of alcohol and/or other drug use (78%), Your impression of the client’s capacity for self-control (70%), and Whether or not the client has informed anyone else such as family or friends about the issue (66%). A quarter of respondents also took the opportunity to provide additional free text responses, which included the following aspects (in no particular order): client insight, social support, expressed intent to commit abuse, duration of the attraction, the presence of alternative (nonabusive) opportunities to gratify sexual needs, use of child sexual abuse images, ego syntonic versus dystonic nature of interest, general lifestyle stability, cooperation with safety plan and continued attendance at therapy, therapist sense of knowing the client well, general criminal history, mental health, indicators of increasing acute risk such as spending time with children or grooming behaviors, and client wishes regarding confidentiality.

Threshold estimates

First, the methodological assumption that the 20 hypothetical scenarios intended to vary on risk were actually perceived as such by participants and in the direction intended was evaluated. This was supported, with the average proportion of respondents opting to disclose increasing linearly as the number of included risk elements increased: for the baseline scenario of a client seeking help with none of the four elements present (i.e., no specific target of attraction, not living with children, no other foreseeable access to children, and evidence of self-control capacity), no respondents (0%) indicated they would opt to disclose client information to authorities. For the four scenarios with one risk element, on average, 11% opted to disclose. This increased to 31% for scenarios with two risk elements present (in varying combinations), 64% for three risk elements, and 84% when all four elements were present (i.e., there was a specific target, the client did have foreseeable access, the client lived with children, and there was evidence of self-control problems).

With that assumption confirmed, the next step was to proceed with exploring thresholds for disclosing client information in the context of varying risk, using the summed 12-point scale estimating each participant’s threshold relative to others. Sample scores on this scale exhibited the full possible range, from 0 (n = 1; 1%) to 12 (n = 11; 10%), with higher scores indicating a higher threshold for opting to make a disclosure. The mean score was 7.10 (SD = 3.07), and the median was 7, indicating that the average participant would have chosen to maintain confidentiality rather than disclose for around seven out of the 12 scenarios presented to them, or a little more than half the time. This slight negative skew (in the direction of maintaining confidentiality) mirrored the distribution found on the self-assessed threshold variable, depicted in Figure 5, but was less pronounced. Data were bimodal, at 6 and 8 (n = 13; 14% each). Figure 6 displays the full distribution for estimated thresholds, which as can be seen is relatively flat indicating broadly varying thresholds across the sample.

Figure 6.

Distribution for estimated thresholds variable (N = 112).

Were Thresholds Associated With Demographic or Professional Variables?

Each of the recorded demographic variables (gender, ethnicity, age) and professional variables (profession, years of professional experience, and specific relevant experience, that is, whether or not the participant had experience within their professional practice of a client seeking help in relation sexual interest in children) were analyzed, to explore for any associations with self-assessed or estimated thresholds. Nonparametric tests were selected due to preliminary analyses failing to support the assumption of normality.

Regarding gender, Mann–Whitney U tests found no significant differences for either self-assessed (U = 735.00, p = .80) or estimated thresholds (U = 815.50, p = .294). Similarly, Kruskal–Wallis analyses found no differences based on ethnicity (χ²(4) = 8.631, p = .071 for self-assessed; χ²(4) = 6.63, p = .157 for estimated thresholds), or profession (χ²(3) = 1.75, p = .626 for self-assessed; χ²(3) = 1.26, p = .739 for estimated). Spearman’s correlations were nonsignificant between both threshold measures with age (r_S = .07 p = .492 for self-assessed; r_S = .13 p = .158 for estimated) and years of professional experience (r_S = .05 p = .571 for self-assessed; r_S = .12 p = .212 for estimated). Finally, neither threshold measure was found to differ on the basis of specific relevant experience (U = 1,449.59, p = .969 for self-assessed; U = 1,248.50, p = .209 for estimated thresholds). In summary, none of the demographic or professional variables were found to be associated with either self-assessed or estimated thresholds based on hypothetical responses.

Were Thresholds Associated With Comfort, Confidence, or Knowledge?

Most of these associations could be explored using correlational analyses, given they involved ordinal variables (the exception was knowledge accuracy; see below). Again, nonparametric tests were selected due to preliminary analyses indicating nonnormality.

Neither comfort, nor confidence, nor self-assessed knowledge was significantly correlated with self-assessed thresholds; results were as follows: r_S = .06, p = .532 for comfort; r_S = .15, p = .110 for confidence in session; r_S = −.03, p = .795 for confidence if given time to consult; and r_S = −.09, p = .333 for self-assessed knowledge. The same pattern of nonsignificant results was found for estimated thresholds: r_S = .09, p = .374 for comfort; r_S = .07, p = .464 for confidence in session; r_S = −.06, p = .537 for confidence if given time to consult; and r_S = .07, p = .471 for self-assessed knowledge.

Mann–Whitney analyses were carried out to assess for associations between thresholds and the dichotomous knowledge accuracy variables (i.e., whether significant mean differences in the threshold measures were apparent between those who answered each of the four knowledge questions correctly or leaned that way vs. incorrectly). For Questions 3a and 3b, regarding the existence of mandated reporting laws, accurate knowledge (i.e., that reporting was permitted but not mandated) was associated with higher estimated thresholds for disclosure: U = 877.50, p = .044, for 3a; U = 669.00, p = .010, for 3b. For 3b, there was also a significant difference in self-assessed threshold (U = 663.00, p = .006). For ease of interpreting the differences with reference to the threshold scales, comparative means will be reported here as opposed to mean ranks (though it is mean ranks that are employed in Mann–Whitney analyses). Those who got 3a correct had a mean estimated threshold of 8.14 (SD = 2.46) on the 12-point scale, compared with a mean of 6.75 (SD = 3.19) for those who had gotten 3a wrong, believing there to be mandated reporting. For 3b, the corresponding means were 8.61 (SD = 2.39) and 6.71 (SD = 3.12) for estimated threshold and 5.09 (SD = 0.85) and 4.42 (SD = 0.963) for self-assessed threshold. The significant differences in mean ranks equated to small effect sizes for all above-reported analyses (ranging between r = −.19 and r = −.26).

Questions 4a and 4b regarded the existence or otherwise under law of protection from disciplinary proceedings for making or not making a permitted disclosure, respectively. For 4a, a significant difference was found between those responding correctly versus incorrectly in terms of estimated threshold (U = 852.50, p = .044). However, the pattern was in the opposite direction to that seen for Questions 3a and 3b: Those responding correctly (i.e., those aware that they would be protected professionally if they were to make a good faith permitted disclosure) had lower thresholds (M = 6.74, SD = 3.18) than those responding incorrectly (M = 8.22, SD = 2.42); the corresponding effect size was in the small range (r = −.19). No differences were found for Question 4b (U = 741.50, p = .568 for self-assessed thresholds; U = 758.50, p = .690 for estimated thresholds).

Finally, a small to moderate positive correlation was found between the two threshold variables (self-assessed and estimated; r_S = .29, p = .002).

Discussion

This study sought to explore the viability of Dunkelfeld-style preventive treatment in discretionary reporting contexts, by investigating health professionals’ comfort, confidence, and thresholds for disclosure in relation to clients experiencing sexual interest in children. Two specific predictions were additionally evaluated: that professionals would experience uncertainty regarding their relevant legal and ethical duties, and that they would display a bias toward disclosing client confidences to authorities when permitted to.

The major findings can be summarized as follows: The health professionals surveyed tended to report comfort and confidence with the scenario involving a client seeking help regarding sexual interest in children. The majority also rated their knowledge of the relevant legal and ethical duties as sufficient, as opposed to lacking. However, findings were mixed in terms of knowledge accuracy. Concerningly, the vast majority of the sample appeared to believe, inaccurately, that mandated reporting laws were in operation in their jurisdiction (NZ). However, most displayed accurate knowledge regarding the differential protection under law from disciplinary proceedings for good faith decisions to disclose client information, as opposed to good faith decisions to not disclose, suggesting the salience of professional liability concerns among health professionals.

Of particular interest for the purposes of this study were respondents’ thresholds for opting to make a disclosure when permitted. Overall, the sample self-reported a general inclination toward the option of maintaining client confidentiality, as opposed to breaching this by disclosing perceived potential risks to authorities when permitted. Estimated thresholds (based on hypothetical responding) exhibited broader variance, but were correlated with self-reported threshold and similarly reflected a general tendency in the direction of confidentiality. However, this was only slight, with participants on average opting to maintain confidentiality in just more than half of scenarios presented to them. Threshold variance (self-assessed or estimated) was not associated with demographic or professional variables.

The overall trends toward comfort, confidence, and sufficient self-assessed knowledge countered the first prediction, of uncertainty among health professionals. Regarding knowledge accuracy, however, the high proportion who incorrectly believed (or guessed) that reporting of risk was mandated under NZ law can conversely be viewed as supporting the prediction—despite the sample tending to think that their knowledge was sufficient, it was in fact overwhelmingly inaccurate on this key aspect. Furthermore, despite the general trends, responses did vary, with a substantial minority (11%-38%) reporting some degree of discomfort, lack of confidence, and/or the sense that their knowledge was lacking. The second prediction, of a bias in favor of making a report to authorities when permitted, was not generally supported. Both the self-assessed thresholds of the sample and the estimated thresholds based on hypothetical responding indicated the opposite tendency: a preference toward maintaining confidentiality. However, this tendency was relatively slight for estimated thresholds, for which the sample displayed broad variation. Contrary to the predicted general bias toward disclosure, any biases or inclinations with regard to reporting appear to vary more at the individual rather than general level. Indeed, findings regarding both self-assessed and estimated thresholds support the likelihood that some health professionals lean toward disclosing, while others are inclined in the opposite direction.

An overarching goal of this study was to explore the potential viability of preventive treatment services in the style of Germany’s Prevention Project Dunkelfeld, in jurisdictions with provisions for discretionary (but not mandated) reporting. Implications of the current findings in relation to viability will now be considered, along with implications for how viability might be able to be enhanced. First, it is clear that there are many professionals who not only feel a sense of comfort and confidence in responding to disclosures of pedophilic interest from help-seeking clients, but also self-report and demonstrate an inclination toward keeping such information confidential even when reporting it on to authorities would be permitted and safer for them professionally. To the extent that the current sample is representative, such a stance even seems in the majority. However, problematically for potential clients who are in need and motivated to seek help, individual professionals’ thresholds or inclinations appear rather difficult to anticipate, with neither demographic (e.g., age, gender) nor professional variables (e.g., profession) bearing any association. This situation could prove too risky for potential clients to come forward (since although the professional they select to confide in might not report them to authorities, they also might). As such, the key barrier to prehabilitation viability in a mandated reporting environment may remain prohibitive in a discretionary one. Remember also that the broader scope for permitted disclosures in NZ may in fact exacerbate this barrier even more relative to many mandated reporting environments, by virtue of including situations of mere perceived risk, as opposed to knowledge or suspicion of actual abuse. However, the ability to exercise discretion, viewed alongside current findings regarding professionals’ general inclination toward maintaining confidentiality, provides at minimum a promising starting point for prehabilitation viability. In the author’s view, results are encouraging enough to suggest that a purpose-designed service, employing staff who concur with the rationale and potential value of prehabilitation and tend toward higher thresholds, operating well within relevant laws and ethical principles with a clear and accessible policy regarding any circumstances under which permitted disclosures will be made, could be viable.

Although demographics did not, one variable that did bear an association with disclosure thresholds was the accuracy of participants’ knowledge of the relevant legal context. Those who understood that mandated reporting was not a feature of their jurisdiction demonstrated a higher threshold for opting to make a permitted disclosure (this was despite accurate information, i.e., that disclosure was permitted but not mandated, being provided prior to the threshold assessments, thereby ruling out the possibility that decisions to disclose were due to the mistaken understanding of a lack of choice). This finding provides one potential avenue for further enhancing the viability of prehabilitation: targeted information provision to increase awareness that reporting of perceived client risk is not mandated. Further support for this suggestion comes from the high proportion of those reporting low comfort or confidence with the scenario emphasizing uncertainty about relevant legal obligations as a reason. Addressing this identified training need could further enhance clinicians’ sense of assuredness in dealing with a scenario of this nature, as well as perhaps encourage more consideration of the option of exercising discretion to not make a permitted disclosure. Undertaking such training could contribute to continuing professional development programs typically required by regulatory bodies in NZ for annual practicing certificate provision under the Health Practitioner Competence Assurance Act 2003.

Interestingly, an entirely different pattern of association with thresholds was found for knowledge accuracy items relating to protection from disciplinary proceedings with respect to permitted disclosures. Those who displayed an accurate awareness (or guessed correctly) that good faith disclosures of risk to children are protected from professional disciplinary proceedings, had significantly lower estimated thresholds for making a report than their counterparts who answered incorrectly. This suggests that, although support was not found for a bias toward disclosure in general, the legislative clause in Question 4a (see caption beneath Figure 4) may promote a level of risk aversion for those who are aware of it. Additional findings may be viewed as supporting this suggestion: Protecting myself professionally was one of the top three reasons endorsed for participants’ self-reported inclination toward reporting risk to authorities, endorsed by half of those inclined in that direction (however, it is noted that the corresponding number of respondents was small, n = 8, due to the majority of the sample being inclined in the opposite direction, i.e., to maintain confidentiality). Conversely, Protecting myself professionally was not highly endorsed by those whose self-reported inclination was toward maintaining client confidentiality: it was cited by only 8% or five of the n = 64 individuals inclined in that direction. Potential avenues to enhance prehabilitation viability based on these results are unclear. Gaining an evidence-based sense of the scope of the professional risk and liability issues in a given jurisdiction could be useful, for example, by accessing the numbers of actual proceedings against those who exercised their discretion in good faith to not make a permitted disclosure, and the outcomes of any such cases. It may be that the risk, though understandably salient for professionals, could actually be considered negligible with such information brought to light. Opening discussions with professional registration bodies and/or disciplinary panelists regarding the rationale for and value of prehabilitation, the imperativeness of client confidence to its viability, and how these aspects may legitimately weigh into discretionary decisions regarding permitted disclosures may also be worthwhile.

Although the current study was intended to have relevance for jurisdictions beyond the immediate study setting (NZ), particularly those that are similar in having discretionary but not mandated reporting laws in effect, the generalizability of findings should not be assumed. Subsequent research wishing to consider prehabilitation viability in other jurisdictions could use the current design as a model, but should substitute the survey content to align with local laws. As a further limitation to note, the opt-in design of the survey may have limited the representativeness of the sample in ways that may have impacted on the results; for example, professionals more interested in and favorable to prevention concepts may have been more inclined to participate, potentially skewing results. Future research could consider alternative recruitment strategies for this reason, such as targeting random samples via professional registration bodies. In addition, the diversity of the sample in terms of gender and ethnic background, though consistent with that of the targeted health professions in NZ as noted,² does leave much to be desired. In particular, males and Māori were underrepresented relative to the general population. Given the contrasting overrepresentation of these groups in samples convicted for the very behaviors prehabilitation aims to prevent (e.g., see Beggs & Grace, 2011), this limited diversity in terms of available professionals for those in need to approach and be willing to place trust in could pose a further barrier to viability. Finally, although the design of the survey reflected the exploratory nature of the current study, associated limitations of measurement should also be addressed in future research. An example is the summed 12-point threshold scale derived from participants’ dichotomous forced-choice responses to scenarios designed to vary on harm likelihood. The interpretation of this scale carried a number of untested assumptions, including that each of the harm-likelihood factors (i.e., existence of a specific target of attraction, children living in the household, unsupervised contact with children, and self-regulation capacity) should be weighted equally. Future research should examine the support for this measurement approach, for instance, by exploring the relative weighting of the 20 scenarios, as well as the overall factor structure of this measure. Keeping these limitations in mind, tentative conclusions can be drawn from this exploratory study and are summarized below.

This study has explored the viability of selected primary prevention or prehabilitation services for individuals with pedophilic interests, in nonmandated but discretionary reporting environments. Results suggested varying and unpredictable personal inclinations and thresholds for reporting of risk information to authorities among health professionals, which could operate as a barrier just as they do in mandated reporting environments (and potentially even more so compared with settings that mandate only knowledge or suspicion of actual abuse, but not perceived risk). However, general trends among professionals were revealed to tend toward comfort, confidence, and the inclination to maintain client confidentiality. Purpose-designed prehabilitation services with clear accessible policies and well-informed supportive staff could be a promising way forward in addressing this barrier. Provision of clear and accurate information regarding the legal context was identified as a key training need, the addressing of which could further enhance prehabilitation viability.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Sarah M. Beggs Christofferson

Notes

Author Biography

Sarah M. Beggs Christofferson, PhD, PGDipClinPsyc, is a senior lecturer in clinical psychology at the University of Canterbury, New Zealand; chair of the Institute of Criminal Justice and Forensic Psychology; and a registered clinical psychologist. Her research interests include assessment and treatment of sexual offending behavior, assessing therapeutic change, and sexual abuse prevention.

References

Alanko

Salo

Mokros

Santtila

(2013). Evidence for heritability of adult men’s sexual interest in youth under age 16 from a population-based extended twin design. The Journal of Sexual Medicine, 10, 1090-1099. doi:10.1111/jsm.12067

Australian Institute of Family Studies. (2017). Child Family Community Australia (CFCA) resource sheet: Mandatory reporting of child abuse and neglect. Retrieved from https://aifs.gov.au/cfca/publications/mandatory-reporting-child-abuse-and-neglect

Beauchamp

T. L.

Childress

J. F.

(2013). Principles of biomedical ethics (7th ed.). Oxford, UK: Oxford University Press.

Becker

J. V.

Reilly

D. W.

(1999). Preventing sexual abuse and assault. Sexual Abuse, 11, 267-278. doi:10.1177/107906329901100403

Beggs

S. M.

Grace

R. C.

(2011). Treatment gain for sexual offenders against children predicts reduced recidivism: A comparative validity study. Journal of Consulting and Clinical Psychology, 79, 182-192. doi:10.1037/a0022900

Beier

K. M.

Ahlers

C. J.

Goecker

Neutze

Mundt

I. A.

Hupp

Schaefer

G. A.

(2009). Can pedophiles be reached for primary prevention of child sexual abuse? First results of the Berlin Prevention Project Dunkelfeld (PPD). The Journal of Forensic Psychiatry & Psychology, 20, 851-867. doi:10.1080/14789940903174188

Beier

K. M.

Grundmann

Kuhle

L. F.

Scherner

Konrad

Amelung

(2015). The German dunkelfeld project: A pilot study to prevent child sexual abuse and the use of child abusive images. The Journal of Sexual Medicine, 12, 529-542. doi:10.1111/jsm.12785

Beier

K. M.

Neutze

Mundt

I. A.

Ahlers

C. J.

Goecker

Konrad

Schaefer

G. A.

(2009). Encouraging self-identified pedophiles and hebephiles to seek professional help: First results of the Prevention Project Dunkelfeld (PPD). Child Abuse & Neglect, 33, 545-549. doi:10.1080/14789940903174188

Centers for Disease Control and Prevention. (2004). Sexual violence prevention: Beginning the dialogue. Retrieved from https://www.cdc.gov/violenceprevention/pdf/svprevention-a.pdf

10.

Chasan-Taber

Tabachnick

(1999). Evaluation of a child sexual abuse prevention program. Sexual abuse: A journal of research and treatment, 11, 279-292.

11.

Child Welfare Information Gateway. (2016). Mandatory reporters of child abuse and neglect. Washington, DC: U.S. Department of Health and Human Services, Children’s Bureau. Retrieved from https://www.childwelfare.gov/topics/systemwide/laws-policies/statutes/manda/

12.

Crimes Act. (1961). Retrieved from http://www.legislation.govt.nz/act/public/1961/0043/137.0/DLM327382.html

13.

Ditmyer

M. M.

Topp

Pifer

(2002). Prehabilitation in preparation for orthopaedic surgery. Orthopaedic Nursing, 21, 43-54. doi:10.1097/00006416-200209000-00008

14.

Gannon

T. A.

Olver

M. E.

Mallion

J. S.

James

(2019). Does specialized psychological treatment for offending reduce recidivism? A meta-analysis examining staff and program variables as predictors of treatment effectiveness. Clinical Psychology Review. Advanced online publication. doi:10.1016/j.cpr.2019.101752

15.

Harris

(2013, October). Do New Zealand psychiatrists have a duty to protect potential victims of their patients’ violence? A dissertation submitted in partial fulfilment of the degree of Bachelor of Laws (Honours) at the University of Otago, Dunedin. Retrieved from https://www.otago.ac.nz/law/research/journals/otago065266.pdf

16.

Health Information Privacy Code. (1994). Retrieved from https://privacy.org.nz/assets/Files/Codes-of-Practice-materials/Consolidated-HIPC-current-as-of-28-Sept-17.pdf

17.

Health Practitioners Competence Assurance Act. (2003). Retrieved from http://www.legislation.govt.nz/act/public/2003/0048/latest/DLM203312.html

18.

Jahnke

Hoyer

(2013). Stigmatization of people with pedophilia: A blind spot in stigma research. International Journal of Sexual Health, 25, 169-184. doi:10.1080/19317611.2013.795921

19.

Lasher

M. P.

Stinson

J. D.

(2017). Adults with pedophilic interests in the United States: Current practices and suggestions for future policy and research. Archives of Sexual Behavior, 46, 659-670. doi:10.1007/s10508-016-0822-3

20.

Leclerc

Chiu

Y. N.

Cale

(2016). Sexual violence and abuse against children: A first review through the lens of environmental criminology. International Journal of Offender Therapy and Comparative Criminology, 60, 743-765. doi:10.1177/0306624X14564319

21.

Letourneau

E. J.

Brown

D. S.

Fang

Hassan

Mercy

J. A.

(2018). The economic burden of child sexual abuse in the United States. Child Abuse & Neglect, 79, 413-422. doi:10.1016/j.chiabu.2018.02.020

22.

Levenson

J. S.

Willis

G. M.

Vicencio

C. P.

(2017). Obstacles to help-seeking for sexual offenders: Implications for prevention of sexual abuse. Journal of Child Sexual Abuse, 26, 99-120. doi:10.1080/10538712.2016.1276116

23.

Lloyd

C. D.

Hanby

L. J.

Serin

R. C.

(2014). Rehabilitation group coparticipants’ risk levels are associated with offenders’ treatment performance, treatment change, and recidivism. Journal of Consulting and Clinical Psychology, 82, 298-311. doi:10.1037/a0035360

24.

Maniglio

(2009). The impact of child sexual abuse on health: A systematic review of reviews. Clinical Psychology Review, 29, 647-657. doi:10.1016/j.cpr.2009.08.003

25.

Mann

R. E.

Hanson

R. K.

Thornton

(2010). Assessing risk for sexual recidivism: Some proposals on the nature of psychologically meaningful risk factors. Sexual Abuse, 22, 191-217. doi:10.1177/1079063210366039

26.

McCartan

(2004). “Here there be monsters”: The public’s perception of paedophiles with particular reference to Belfast and Leicester. Medicine, Science and the Law, 44, 327-342. doi:10.1258/rsmmsl.44.4.327

27.

Ministry of Health. (2009). Psychologist workforce: Summary results from the 2009 health workforce annual survey. Retrieved from http://www.psychologistsboard.org.nz/cms_show_download.php?id=87

28.

New Zealand Ministry of Justice. (2016). Number of people convicted of sexual offences against a child or young person (aged 16 years or younger) in 2015, by offending history. Unpublished dataset released to S. Christofferson.

29.

New Zealand Psychologists’ Board. (2002). Code of ethics for psychologists working in aotearoa/New Zealand. Wellington: Author.

30.

Oranga Tamariki Act. (1989). Retrieved from http://www.legislation.govt.nz/act/public/1989/0024/121.0/DLM147088.html

31.

Schaefer

G. A.

Mundt

I. A.

Feelgood

Hupp

Neutze

Ahlers

C. J.

. . . Beier

K. M.

(2010). Potential and Dunkelfeld offenders: Two neglected target groups for prevention of child sexual abuse. International Journal of Law and Psychiatry, 33, 154-163. doi:10.1016/j.ijlp.2010.03.005

32.

Stats NZ (n.d.). 2013 Census QuickStats about culture and identity. Retrieved from http://archive.stats.govt.nz/Census/2013-census/profile-and-summary-reports/quickstats-culture-identity/ethnic-groups-NZ.aspx

33.

Tarasoff v. Regents of the University of California. (1976). 131 Cal. Rptr. 14 (Cal. 1976).