Therapeutic Engagement and Treatment Progress: Developing and Testing an In-Treatment Measure of Client Engagement Among Sex Offenders in a Group Program

Abstract

The engagement process of sexual offenders in group-based cognitive-behavioral treatment is an important area of study. Disclosure management style (DMS), a model developed from grounded-theory research of men undertaking a prototypical program, provides a framework to assess engagement in treatment. Our goal was to develop a quantitative measure of DMS, to test its reliability and validity, and to evaluate its utility as a measure of treatment progress by examining relationships between DMS and established measures of treatment change. We studied a sample of men (N = 93) who undertook an intensive prison-based treatment program in New Zealand. Variables included DMS measures, psychometric measures of dynamic risk and treatment change, static risk, clinician rating of treatment progress, and recidivism outcomes. We found that (a) DMS shows an acceptable level of agreement between independent raters, (b) DMS-based ratings of engagement changed over the course of treatment and were correlated with measures of change based on offender self-reports, and (c) offenders showed heterogeneity in terms of their trajectories of change as assessed by DMS ratings.

Keywords

sexual offender treatment treatment progress responsivity child sexual abuse recidivism engagement

Introduction

Group-based cognitive-behavioral programs represent an “industry standard” in sexual offender treatment (McGrath, Cumming, Burchard, Zeoli, & Ellerby, 2010; McGrath, Hoke, & Vojtisek, 1998). Their goal is to target those factors that are both empirically related to recidivism (Beggs, 2010) and potentially amenable to change, such as personal beliefs and attitudes, social and emotional functioning, and sexual deviance (Craissati & Beech, 2003). There is considerable evidence supporting the effectiveness of such programs in reducing reoffending (see Kim, Benekos, & Merlo, 2016), and this has been taken as support for these targets as having a causal role in offending.

Appropriateness of treatment targets aside, the nature and intensity of offenders’ participation in programs is clearly critical to effectiveness. In this article, we use the term engagement to refer to the offender’s commitment to the goals and processes of treatment, demonstrated by a comprehensive, sustained, and proactive style of participation. The engagement of sex offenders, however, is notoriously problematic and attention in the literature has now begun in earnest to consider those features of treatment context and process that (as “responsivity factors”; see Polaschek, 2012) influence engagement (e.g., Beech & Hamilton-Giachritsis, 2005; Harkins & Beech, 2007; Marshall & Burton, 2010; Serran & Marshall, 2010; Ware, 2011). The influential work of Andrews and Bonta (2010) identifies responsivity, along with risk and need, as a core principle associated with effective criminal justice intervention. Responsivity is concerned with both the activities of the treatment provider and the attributes brought by the client. In practice, attention to it involves tailoring and delivering services in ways that are most likely to promote uptake of program content. The critical importance of engagement to the success of offender programs (irrespective of program type, criminogenic need, or treatment setting) has also been established directly (Drieschner & Verschuur, 2010; McMurran & Ward, 2010; Scott & King, 2007). Research into engagement then represents an important aspect of more general process-related developments in the offender treatment literature (see Holdsworth, Bowen, Brown, & Howat, 2014, for a review).

One program factor that might act as a barrier to engagement is the requirement, in most sex offender treatment programs, of a high level of personal disclosure (Walji, Simpson, & Weatherhead, 2013). The term disclosure is used here in a broad sense and not merely with respect to, say, the “confession” of offense details. Rather it refers to the propensity to participate in open, honest, and direct exchange, confronting very challenging matters in relation to one’s inner life and exploitive intentions. Such disclosure is intended to reveal actor intentions and conduct, so that these might be examined and modified in the group setting. Disclosure expectations are common to a broad range of sex offender program types, including those influenced by strengths-based approaches as well as those that emphasize a more conventional cognitive-behavioral therapy (CBT)-driven, relapse-prevention orientation. Nevertheless, these expectations are very challenging to participants. Valid participation requires attention to deeply personal matters such as sexual arousal, intimacy, emotional regulation, self-evaluation, and victim empathy. These are factors that have the potential to elicit potentially distressing emotions, such as shame, humiliation, fear, rejection; as well as social alienation and psychological isolation. Particularly in a group setting, the intensity of these painful feelings is likely to be high (see also Ware & Mann, 2012), and it is not surprising that disclosure expectations can test participant engagement.

Disclosure management style (DMS; Frost, 2004; Frost & Connolly, 2004; Frost, Daniels, & Hudson, 2006) refers to how participants in treatment handle disclosure expectations. The DMS model was developed from grounded-theory research into the engagement strategies of men undertaking a prototypical program (see Table 1 for a list of constituent treatment modules). It assesses participants’ strategies toward the disclosure of offense-related personal information (Barrett, Wilson, & Long, 2003; Holdsworth et al., 2014). Previous research has indicated that, faced with the requirement of disclosure, the way in which participants respond differs along the dimensions of openness and self-directedness, which together define their DMS. We propose that this construct has utility as a measure of treatment engagement and is likely to influence treatment progress. A primary focus on DMS as an indicator of engagement seems justified by the centrality of disclosure in the treatment process, discussed above, as well as both validity considerations and the pragmatics of measurement with such an elusive construct (see Holdsworth et al., 2014). Its utility would be supported if it was found to be related to indicators of treatment progress linked to program content. Accordingly, the current study tests the reliability and validity of a quantitative measure of DMS in men who had sexually offended against children and were enrolled in an in-prison treatment program.

Table 1.

Kia Marama Program Structure and DMS Assessments.

Week	Content	DMS assessment
1-2	Clinical assessment	“Time 1”
3-4	Orientation to group therapy
5-11	Offense mapping	“Time 2”
12-13	Sexual arousal reconditioning
14-17	Victim impact and empathy	“Time 3”
18-21	Relationship and sexuality
22-25	Mood management
26-29	Release planning
30-33	Clinical reassessment	“Time 4”

Note. DMS = disclosure management style.

The remainder of this article is organized as follows. First, we review the background and development of the DMS construct and propose a quantitative measure of DMS. Next, we present the results of an empirical study that evaluates the reliability and validity of this DMS measure, and explore the relationship between DMS and measures of within-treatment change that have been shown to be predictive of recidivism (Beggs & Grace, 2011). Finally, in the discussion, we consider implications for both clinical practice and risk assessment, limitations of the study, and directions for future research.

DMS

DMS was developed in a study of the reactions of group-treatment participants to a challenging, early component of treatment. The component, often called the offense chain, requires participants to give an account of the sequence of cognitive, affective, and behavioral events that culminated in their offending, thereby identifying as someone who has offended sexually against children. Their account is then analyzed and elaborated by the group in a therapist-led discussion of the offender’s account. For the purposes of the study, immediately afterward, participants recorded the most salient experiences in this session, and next, in an individual interview with the researcher, articulated these experiences in a moment-by-moment analysis of each. Details of the selected experiences were then subjected to a grounded-theory analysis (Frost, 2004).

The emergence of the DMS model was a significant outcome from this study. The model suggests that when presenting their offense chains, men take up one of four relatively distinct orientations to the challenges of the disclosure encounter, labeled as Exploratory (relatively proactive, direct, open, and reflective), Oppositional (generally explicit attempts to control disclosure and to dismiss or close down challenge), Evasive (seeking to avoid experiencing personal exposure), and Placatory (approach-laden efforts to win interpersonal acceptability). According to the DMS model, these reflect positions on two dimensions represented in Figure 1. The x-axis represents the man’s strategy with respect to his attempts to influence the flow of communication in the group, varying between extremes of open and closed. The y-axis indicates the degree to which the man privileges either self (his own needs) or other (his perception of others’ expectations of him) as the primary point of reference regarding the achievement of his immediate goals in the context of the group. High self-directedness (a positive y value) is indicated by a conscious realization of self-generated and self-referenced priorities in the encounter. An open interpersonal strategy (a positive x value) is indicated by an inclination to promote the sharing of therapeutic resources through openness to ideas, suggestions, advice, and critique. For example, in the case of a man who is oriented to self-referenced goals (e.g., “I need to stay focused on my need for clarity here”), combined with an open strategy (e.g., “The best way to meet my goals in this setting is to encourage others to give their opinion freely”), it results in an exploratory DMS. On the contrary, a man who is inclined to give primary consideration to the concerns of others (e.g., “They are not going to like that if I bring it up”) and seeks to prevent open exchange (e.g., “If I let him keep talking in this way about me, the others are going to realize how disgusting I was”) indicates an evasive DMS.

Figure 1.

Disclosure management styles.

With respect to the program goals and requirements for achieving them, Frost and Connolly (2004) argued that only the qualities of relatively uninhibited, discovery-driven enquiry and self-referenced motivation (Exploratory style) meet the requirements of any process that relies heavily upon a clinical culture of active collaboration and responsiveness to interpersonal feedback. The other three styles imply the presence of immediate concerns of the participant that are, at best, distracting and, at worst, in direct opposition to the ultimate treatment goals. Thus, we expected that measures of progress toward an exploratory DMS would be positively correlated with other (content goal-related) measures, for example, prosocial changes on dynamic risk factors (Beggs & Grace, 2011).

Goal of the Current Research

Although earlier work has suggested that the DMS construct is a useful addition to assessment and clinical practice, the construct itself has not yet been subject to empirical testing. Thus, the goal of the current research was to investigate the reliability of the DMS construct, to explore its coherence and integrity over the course of treatment, and to examine relationships between DMS, (psychometric) measures of within-treatment change, and risk for reoffending. Specifically, we planned to develop and validate a graphical rating scale intended to provide quantitative measures of the two DMS dimensions in Figure 1. Our major questions concerned (a) whether DMS measure was sufficiently well operationalized that independent raters would show an acceptable level of agreement about the DMS of individual offenders; (b) whether DMS ratings would show convergent validity with measures of treatment change based on self-reports (Beggs & Grace, 2011), which would indicate that progression toward an exploratory style could be viewed as an adjunct to successful treatment; and (c) whether offenders showed heterogeneity in terms of their changes in DMS rating over treatment. We also planned to test whether progression toward an exploratory DMS would be associated with a reduced likelihood of sexual recidivism, accounting for pretreatment levels of dynamic and static risk although this analysis was considered preliminary because of the limited sample size.

Method

Participants and Program

The participants were 93 men who had attended the Kia Marama Special Treatment Unit (Rolleston Prison, Christchurch, New Zealand) for child sexual offenders and who were released from prison between July 2001 and June 2007. All gave written consent for their file information to be used for research purposes; ethical approval was obtained from the University of Canterbury Human Ethics Committee.

The men were aged between 19.3 and 76.9 years at release, with an average age of 43.32 (SD = 12.29). Ethnically, most (65.6%) were New Zealand European, while 20.4% were Māori or Pacific Island, and 14.0% were other or unspecified. Sixty-one (65.6%) were intrafamilial offenders whose victims came entirely from their own immediate or extended families, whereas 32 (34.4%) were extrafamilial offenders with at least one victim who was unrelated to them. All but five men successfully completed the program (94.6%; 88/93). During the 6-year (average) follow-up period, 6.5% (6/93) were convicted of a new sexual offense, 5.4% (5/93) were convicted of a new violent offense, and 28.0% (26/93) were convicted of a new general (i.e., nonsexual and nonviolent) offense. The average failure times were 576.6, 818.5, and 957.6 days for sexual, violent, and general recidivists, respectively.

The program is group-based, with a cognitive-behavioral orientation, an emphasis on relapse prevention, and organized as a therapeutic community. Treatment lasts for 33 weeks during which group sessions of eight to 10 men are conducted three times a week (a total of some 250 hr of treatment). Modules include group norm-building, offense chain (offense mapping), arousal reconditioning, victim impact and empathy, mood management, relationship skills, sexuality, and relapse prevention. For more information about the Kia Marama program, see Hudson, Marshall, Ward, Johnston, and Jones (1995) and Hudson, Wales, and Ward (1998).

DMS Measure

Clinicians responsible for facilitating group therapy sessions at Kia Marama were recruited to participate in the research. In total, there were seven therapists who were responsible for rating 12 therapy groups. They were trained in DMS and behavioral indicators of the four styles (Figure 1) by the first author (AF). Specifically, therapists were given a manual that described the DMS model and different behavioral indicators associated with the four disclosure styles. AF provided an overview of the assessment methodology to the therapists as a group and subsequently met individually with each therapist to ensure adequacy of understanding and competence in assessment. Therapists were asked to identify the primary DMS of the men in their groups and to monitor developments with them so that potential changes in DMS for each individual at critical points in the program could be assessed.

The therapists assessed DMS for each man in their group at up to 4 points in treatment (Table 1): Assessment (Program Week 4), offense chain (Week 11), victim empathy (Week 17), and reassessment (Weeks 36-39).

Figure 2 shows how different DMS assessments might be represented in an assessor’s report. DMS was notated by marking a single point such that the X and Y coordinates of the point corresponded to the man’s current position on open-closed and self-other dimensions, respectively (each assessment consisted of one such point).Concurrently but independently, the first author (AF) observed group sessions using live two-way mirror screening or video records, and assessed DMS at the same point in treatment.

Figure 2.

Illustrative plot showing nine different DMS assessments on the two dimensions of DMS.

Quantitative DMS ratings were obtained by first superimposing, over the two-dimensional graph a grid with coordinates from −40 to +40 on each dimension (and with 0,0 at the center). Numerical values for openness and self-directedness were thus taken directly from these plots. To measure change using this graphical system, we converted assessments to distances from an “ideal” disclosure style (maximal openness and maximal self-directedness, the X in Figure 2) at different times during treatment. Distances were calculated separately for AF and the seven group therapists, and for cross-therapist validation, correlations were then computed between distances based on AF’s assessments and those based on the group therapists’ assessments.

Psychometric Measures of Dynamic Risk and Treatment Change

All participating program clients completed a battery of psychometric tests at both pre- and posttreatment (i.e., assessment and reassessment phases). The battery has been described in more detail by Allan, Grace, Rutherford, and Hudson (2007), who identified four dynamic risk factors in the battery: Social Inadequacy (F1), Sexual Interests (F2), Anger/Hostility (F3), and Prooffending Attitudes (F4). Subsequently, Beggs and Grace (2011) developed a methodology for assessing treatment change using the battery in which change scores (i.e., posttreatment–pretreatment, calculated so that positive scores indicated prosocial change) were calculated for each variable. Then, variance associated with pretreatment scores was partialled out, and the resulting residual change scores were standardized and averaged for each dynamic risk factor. Beggs and Grace (2011) showed that the standardized residual change scores (RCZ) were positively correlated with measures of change based on the Violence Risk Scale–Sexual Offender version (VRS-SO; Olver, Wong, Nicholaichuk, & Gordon, 2007) and significantly predicted reductions in recidivism after controlling for static and dynamic risk levels at pretreatment.

Static Risk

The Automated Sexual Recidivism Scale (ASRS; Skelton, Riley, Wales, & Vess, 2006) is an actuarial computer-scored risk assessment measure of relevant static (historical) factors, methodically based on the Static-99 (Hanson & Thornton, 2000) using variables sourced from the New Zealand Department of Corrections’ Criminal History database. The Static-99 is probably the most widely used actuarial assessment for risk of sexual recidivism (Storey, Watt, Jackson, & Hart, 2012). Evaluating its predictive validity using the area under the Receiver Operating Characteristic (area under the curve [AUC] of the ROC), its performance commonly falls in the range 0.68 to 0.75, sometimes decreasing as the follow-up period increases (Hanson & Thornton, 2000; Helmus, Thornton, Hanson, & Babchishin, 2012; Reeves, Ogloff, & Simmons, 2017). The seven-variable ASRS has been found to achieve comparable AUC values, at least to New Zealand offenders. In a study of 1,133 released male sex offenders, Skelton and others (2006) reported AUC values of .78, .75, and .70 for 5-, 10-, and 15-year follow-ups, respectively. Across four risk bands, Low to High, observed sexual recidivism rates among New Zealand sex offenders were .08, .12, .24, and .48 after 10 years. These results confirm that this instrument has substantial predictive validity with the sex-offender population from which the present study’s sample were taken. The items measured in the ASRS are number of prior sex convictions, number of prior sentencing dates, any convictions for noncontact sex offenses, index offense of nonsexual violence, prior nonsexual violence charges, any male victim, and age of offender at time of index offence.

Clinician Rating of Treatment Progress

At the end of treatment, clinicians for each group provided an overall binary rating of each man’s progress (successful/unsuccessful). This rating provided some measure of “success” but was not calibrated with any rigor, being a reflection of the individual clinician’s estimation in otherwise unspecified terms.

Recidivism Outcomes

Criminal history information for all men in the sample was obtained from the computer database maintained by the New Zealand Department of Corrections as of July 1, 2008. Convictions for sexual, violent, or general offenses that occurred postrelease were recorded.

Statistical analyses were conducted with SPSS (v. 22). A significance level of .05 was used.

Results

First, we examined whether assessments of DMS were reliable. Because the first author (AF) assessed all cases, we computed the correlations between his assessments and those of group therapists for both Openness and Directedness dimensions. High positive correlations would suggest that the dimensions were sufficiently well operationalized for the therapists to learn how to assess them. However, low correlations could result from either poorly operationalized constructs, inadequate training given to the therapists, or both.

Table 2 shows the correlations between AF’s assessments and those of the seven group therapists (A-G), separately for each dimension. Because of staff turnover, some therapists contributed fewer assessments than others over the duration of the study. In general, correlations were high; only three were less than .50 (for Therapists C and E on Self-Directedness, and Therapist F on Openness). When assessments were pooled, correlations for Openness and Self-directedness were .66 and .70, respectively. These correlations show that there was moderately good agreement between AF and group therapists, suggesting that both DMS dimensions were sufficiently well operationalized and that the therapists were adequately trained.

Table 2.

Correlations Between Individual Therapists and the AF for Scores on Openness and Self-Directedness.

Therapist	N	Openness	Self-directedness
A	8	.87	.74
B	7	.95	.96
C	54	.62	.42
D	55	.77	.86
E	48	.62	.40
F	24	.44	.78
G	28	.77	.78
All	224	.66	.70

Note. AF = first author; N = The number of assessments made by each therapist.

For the remaining analyses, we pooled assessments across individual therapists. Agreement was adequately stable across assessments, but improved somewhat at Time 4 relative to earlier ones. Correlations between therapists and AF at Times 1, 2, 3, and 4, were .65, .65, .51, and .80, respectively, for Openness, and .71, .66, .65, and .76 for Self-Directedness (n ranged from 52 to 63). Correlations between assessments weakened as the temporal gap between them increased. Figure 3 shows the correlation coefficients for various pairs of assessors as a function of the time between individual assessments. Each of the 4 points above 0 in Figure 3 shows the correlation between distance-from-ideal scores from AF and group therapists at one of the stages in treatment (Times 1, 2, 3, or 4). Unfilled squares show the correlations between AF’s assessments at different times (e.g., Time 1 vs. Time 2, Time 2 vs. Time 3, and Time 3 vs. Time 4 all appear above “1,” Time 1 vs. Time 4 appears above “3”). These correlations decrease as the time between assessments increase, as would be expected if there was nonuniform change in disclosure style across participants over the course of treatment. The same is true for correlations between AF’s and group therapists’ assessments (diamonds in Figure 3) and for successive assessments by group therapists (triangles).

Figure 3.

Engagement and treatment progress: Correlations for various pairs of assessors as a function of the time between them.

To determine whether there were significant changes in DMS scores over treatment, we conducted repeated-measures ANOVAs in which time of assessment and therapy groups were factors. To maximize the available sample size (as DMS was not scored for all offenders at all 4 assessment points), we calculated a pair of DMS scores: A “start” score that was defined as the assessment early in treatment (Time 1 or Time 2) with the greatest distance from the ideal DMS style (i.e., exploratory) and an “end” score that was the last assessment available (either Time 3 or Time 4). Therapy group membership was included as a factor to reflect the potential relatedness of observations by the clinician within that group.

For Openness, the effect of assessment time was significant, F(1, 66) = 4.17, p < .05, η_p² = .06. The mean Openness scores for the start and end assessments were 2.27 and −2.80, respectively, indicating a shift toward an open strategy. The effect of Group and the Group × Time interaction were not significant, F(11, 66) = 1.69, p = .10, η_p² = .22 and F(11, 66) = 1.07, p = .40, η_p² = .15, respectively. For Self-Directedness, there was a significant effect of assessment time, F(1, 66) = 10.45, p < .01, η_p² = .14. The average score increased from −3.52 (start) to 4.81 (end), showing progress toward self-orientation. The effects of Group and the interaction were not significant, F(11, 66) = 1.33, p = .23, η_p² = .18 and F(11, 66) = 0.98, p = .44, η_p² = .14, respectively. Overall, participants were rated as showing a moderate degree of improvement in DMS, both in terms of Openness and Self-Directedness.

To summarize, correlations among different assessors for the two dimensions of Openness and Self-Directedness suggest that these constructs were sufficiently well-defined and operationalized, so that group therapists could reach assessments of individual cases that were reasonably well correlated with assessments by the developer. Assessments taken at different times were highly correlated when the time between them was short, but only moderately correlated when separated in time. The decrease in correlations is consistent with the uneven progress that group participants showed in the disclosure style across participants over time.

Next we examined correlations between these disclosure management change scores and measures of treatment change based on psychometric self-reports and clinician ratings, as well as measures of pretreatment dynamic and static risk measures. Results are shown in Table 3. According to clinician’s ratings of overall progress, treatment was judged as “successful” for 30 of 93 participants (32.26%), and successful treatment was positively correlated with change on DMS r = .39 and .37 for Openness and Self-Directedness, respectively, both ps < .001. Correlations between changes in DMS and dynamic risk based on offender self-reports were also generally positive, particularly for Openness. Change in Openness showed positive and statistically significant correlations with change in Social Inadequacy, r = .23, p < .05, Anger/Hostility, r = .28, p < .05, and Prooffending Attitudes, r = .26, p < .05, as well as the overall psychometric change measure, r = .33, p < .01. Thus, measures of change based on clinician’s ratings of DMS were positively related to measures of change derived from offender’s psychometric self-reports, supporting the validity of DMS as a measure of treatment change.

Table 3.

Correlations Between Disclosure Management Style Scores and Other Measures.

	Recidivism	ASRS	Open-chg	Orient-chg	Success	Social inadequacy (F1)	Sexual interests (F2)	Anger/hostility (F3)	Prooffending attitudes (F4)	Overall dev	F1 chz	F2 chz	F3 chz	F4 chz
Recidivism
ASRS	.09
Open-chg	−.08	.16
Orient-chg	.06	.13	.04
Success	.01	.10	.39***	.37***
Social inadequacy (F1)	−.07	.20	−.09	−.13	−.02
Sexual interests (F2)	.23*	−.11	−.02	.05	−.02	.16
Anger/hostility (F3)	−.01	−.01	−.11	−.24*	−.03	.32**	.18
Prooffending attitudes (F4)	−.09	.26*	−.13	−.08	−.13	.56***	.20	.47***
Overall dev	.11	.22	−.19	−.07	−.18	.64***	.51***	.54***	.71***
F1 chz	−.08	−.15	.23*	.10	.23*	−.17	−.01	−.01	−.19	−.17
F2 chz	−.10	−.18	.09	.05	.04	−.01	.02	−.03	.03	−.08	−.10
F3 chz	−.12	.00	.28*	−.03	.23	−.23	−.20	−.20	−.40**	−.26*	.45***	.07
F4 chz	.04	−.12	.26*	.22	.14	−.14	.27*	−.23*	−.18	−.08	.49***	−.03	.22
Overall chz	−.08	−.19	.33**	.15	.24*	−.21	.04	−.19	−.26*	−.22	.76***	.56***	.64***	.63***

Note. ASRS = Automated Sexual Recidivism Scale.

chz = Residual change score.

dev = Deviance.

*p < .05, **p < .01, ***p <.001.

We asked whether offenders showed heterogeneity in their trajectories of DMS change during treatment. Because of the holistic nature of the DMS assessments, we used cluster analysis to identify subgroups of offenders based on the combination of “start” and “end” measures reported above, rather than analyzing change in “start” and “end” measures separately. Data (n = 85) were analyzed with the two-step cluster algorithm in SPSS, which uses the Bayesian information criterion (BIC) to determine the optimal number of clusters. Results indicated a three-cluster solution. The cluster means for Openness and Self-Directedness for the first and last assessments (StartX, StartY EndX, EndY, respectively) are listed in Table 4. The trajectories of change are graphically displayed in Figure 4 by arrows, where the start and end point of each arrow corresponds to the cluster means in Table 4.

Table 4.

Cluster Means for First and Final Assessments.

	Cluster
	1 (n = 43)	2 (n = 14)	3 (n = 28)	F(2, 82)
StartX	3.58	11.91	−10.79	11.17***
StartY	4.64	−10.58	−12.32	10.31***
EndX	−7.21	−21.23	6.32	16.15***
EndY	−2.22	14.77	6.95	7.12**

Figure 4.

Trajectories of change in the DMS space for different groups identified in the cluster analysis.

Cluster 1 began with a relatively neutral DMS (Ms = 3.58 and 4.64 for starting values of Openness and Self-Directedness, respectively) and showed significant improvement on Openness, t(34) = 4.26, p < .001, d = 0.72, but not Self-Directedness, t(34) = 1.76, p = .09, d = 0.30, by the end of treatment (ending Ms = −7.21 and −2.22). Cluster 2 began with an evasive style (Ms = 11.91 and −10.58 for Openness and Self-Directedness) but made significant progress on both dimensions and showed an exploratory style by the end of treatment, Ms = −21.23 and 13.16; t(12) = 11.92, p < .001, d = 3.31 and t(12) = −5.93, p < .001, d = −1.64. By contrast, Cluster 3 began with a placatory style (Ms = −10.79 and −12.32) and moved toward an oppositional style by the end of treatment (Ms = 1.97 and 6.47), with significant changes toward a closed strategy, t(21) = −3.96, p < .001, d = −0.84, and self-directed orientation, t(21) = −6.19, p < .001, d = −1.32. The clusters differed significantly in terms of clinician’s ratings of overall treatment progress. For Cluster 2, 71.4% of cases (10/14) were rated as successful, compared with 30.2% for Cluster 1 (13/43) and 17.9% for Cluster 3 (5/28), χ²(2) = 12.42, p < .01.

Discussion

This article presented the outcomes of the empirical testing of DMS. DMS was the focus for the study because it is an important component of the broader construct of treatment engagement and may even be indicative of participants’ status on that broader construct. The principal goals of the research were (a) assessment of the clinical coherence of DMS, (b) testing its validity as a measure of therapeutic change, and (c) tracking its consistency with respect to other variables seen to be relevant in recidivism prediction. To achieve these objectives, we (a) tested the reliability of the construct in clinical use, (b) considered correlations with other measures of treatment change, and (c) used cluster analysis to assess homogeneity of change trajectories in DMS among participants

Results indicate that

Agreement between independent raters was acceptable, suggesting that DMS is sufficiently well operationalized to be used in a clinical setting;

Overall, measures of DMS changed significantly over the course of treatment and, overall, the direction of change represented progress toward an optimal self-disclosure position;

However, participants showed heterogeneity in terms of their trajectories of DMS change, with some showing a progression toward optimal self-disclosure, whereas others showed little change;

DMS change was positively correlated with changes in dynamic factors (a) that are known to improve with treatment and (b) improvement in which is known to predict lower recidivism. Thus, movement toward a more optimal style of disclosure management was associated with positive treatment change (Beggs & Grace, 2011).

The results outlined above provide some assurance that the DMS construct is coherent, and can be a reliable construct in a clinical setting. Given training, therapists came to use the scales in the same manner as the developer (AF), as evidenced by the correlations between their ratings and his. These correlations were generally high although there were a few that were less than .5. Together, these observations suggest that attention to DMS training with therapists may be important in ensuring reliable portability of the construct to other clinical settings where difficult self-disclosure is a key requirement, and in maintaining a high standard of clinical utilization.

The ultimate measure of treatment effectiveness for offenders is reoffending statistics. In general, however, efforts to test the association between features of treatment provision and reoffending outcomes have proved neither straightforward nor comprehensive. The development of trustworthy within-treatment measures has therefore been identified as an important direction for study in efforts to enhance rehabilitation and to progress the field (Allan et al., 2007; Beggs, 2010; Beggs & Grace, 2011; McGrath et al., 2010). These proximal measures of treatment change are focused on program-content features—core treatment targets (Grady, Brodersen, & Abramson, 2011). In the current research, we asked whether there are other, responsivity-related, dimensions of the change that occurs in the course of offense-related treatment. More specifically, we asked whether participants’ DMSs changed over the course of treatment and whether the changes observed might be important in supporting progress with program content. We found that DMS improved over the course of treatment, with many offenders moving toward a more open, exploratory style of disclosure. While responsivity-related change is not necessarily the same as risk-related change, our study also suggests that it is both correlated with risk change and associated with lower recidivism.

Correlations among successive assessments of DMS were degraded by the amount of time spent in treatment between them. Had all participants shown similar change in their DMS over the course of treatment, these correlations would have remained high; their progressive reduction with increasing temporal separation therefore suggests considerable unevenness in DMS change across participants over the course of treatment. In the absence of concerted, formal efforts to target DMS, this is hardly surprising, however. This unevenness is consistent with the findings from the cluster analysis, which revealed the existence of three identifiable trajectories of change in DMS, each one representing different levels of change toward the ideal (open, self-directed) disclosure style. Although the largest of the three clusters showed relatively strong progress, others showed less; and additional clinical effort may be required with some offenders to achieve more general treatment progress. We take up this point again below.

Clinical Implications

The outcomes from the study add further weight to the argument, underpinned by previous research and clinical experience (see Frost & Connolly, 2004), that DMS has potential as a means of monitoring therapeutic engagement throughout the challenging experience of group-based treatment for sexual offenders. As a reasonably reliable source of such feedback, the construct might also guide establishment and maintenance of therapeutic relationships. With the corroboration of data from this study, treatment providers can now advise offenders in these programs, with more assurance and with less speculation, that an open and self-directed (“exploratory”) approach to personal self-disclosure is likely to enhance the benefit of treatment. Change toward an exploratory style might be facilitated with program participants by, for instance, having them self-evaluate their present position with regard to openness and “ownership,” and by promoting (perhaps through modeling) a spirit of curiosity in their approach to treatment. Such facilitator practices of course revolve around skills that are more to do with the quality of therapeutic relationship processes and less to do with manualized prescription. Given that DMS has been seen to be plausibly assessed pretreatment (Frost & Connolly, 2004), this might be managed in pretreatment work, either in group or individual contexts.

The DMS construct was developed in the context of a sex offender treatment program that was partly derived from a relapse-prevention perspective. Its initial focus on an offense chain places an early demand on participants for frank disclosure about their offending, with a view to putting in place protections against repetition. Other approaches to sex-offender treatment might not have the same starting point, but we have difficulty conceiving of one that does not involve honest and open exchange about intimate matters such as sexual arousal, intimacy, emotional regulation, and self-evaluation, nor one in which the key dimensions of openness and self-directedness are unimportant. Indeed, this latter consideration is arguably critical to applications outside of sex offender treatment. We believe then that the DMS has widespread applicability, not restricted to contexts in which offense disclosure is central. Specifically, DMS may be a useful indicator in proximally monitoring the treatment progress of sexual offenders, by tracking the course of their engagement in treatment (see Holdsworth et al., 2014). If so, it would complement within-treatment measures of treatment progress as an indicator of how well participants engage with program content (i.e., as a responsivity measure).

Monitoring responsivity generally, and DMS in particular, during treatment may be as important as monitoring change in relation to program targets. The reason is that many program participants commence treatment with a DMS position that is neutral (i.e., suboptimal) or worse, and disclosure is a central requirement of valid participation in the program. Thus, improving DMS might be a very significant method of improving engagement, and progress here might often be crucial to achieving progress in relation to program targets. Assigning central concern to the assessed DMS of participants in considering their engagement also helps identify impediments to therapeutic engagement, such as fears around social exposure. Equally, it might also assist by exposing “false positives” such as the desire to please, which often masquerade as treatment engagement. The opportunity presented by DMS as a construct here lies in the fact that an exploratory style indicates a multidimensional positive response to the invitation to engage (by virtue of its attention to the interpersonal elements of openness and directedness), extending well beyond simple compliance. By making the dimensions of DMS explicit, they are placed on the agenda of change targets while participants themselves are sensitized to the phenomena in the group and to their own tendencies in relation to it.

Our results suggest that clients who began treatment at a more or less neutral position (Cluster 1) in the DMS space made some moderate gains in terms of increased level of openness, but otherwise shifted little in this respect. A second cluster (Cluster 3) who commenced treatment with a moderately placatory approach completed with a generally neutral position. While we could speculate optimistically that the effect of treatment in the case of those who present as prioritizing social desirability is to enhance confidence and develop firmer interpersonal boundaries (becoming more self-determined and discreet), taken together, Cluster 1 and Cluster 3 suggest overall movement is toward a more neutral position. Less change in DMS was associated with less successful treatment according to therapist ratings and psychometric test scores, and also with higher recidivism. The most striking trajectory trend according to the data is with regard to Cluster 2. These participants showed a relatively pronounced movement away from an evasive position into an unequivocally exploratory one. This suggests these clients are more likely to benefit from treatment compared with those with neutral (Cluster 1) or closed (Cluster 3) tendencies. We might infer that the therapeutic qualities at work here are those that enhance trust and counter interpersonal anxiety, factors that have been empirically associated with group psychotherapy (Burlingame, Strauss, & Joyce, 2013).

Limitations and Future Research

Although interrater correlations were good overall, it is acknowledged that some were very high, whereas a few were only moderate. This combination suggests that the potential coherence of the construct was not fully realized in training of some of the therapists, and that there is room for improvement here. An additional concern is that therapists in this study had access to considerable expertise with the construct, in that they were trained in DMS assessment by the developer. Together, these observations suggest that additional attention to DMS training may be important in ensuring portability of the construct to other clinical settings and in maintaining a high standard of clinical utilization.

Despite providing the benefits described above, the study was limited by not using a fully nested design. For example, a study in which observations were consistently made by therapists within particular groups would allow for therapist and group effects to be estimated independently and potentially provide stronger evidence of causal effects related to DMS. Because of the demands of the treatment program, such an approach was not practical in the present study.

While our study suggests that responsivity-related change is correlated with treatment change, the present results should be regarded as preliminary, given the very limited sample size and the number of variables entered into the equation. The promise generated from this research is sufficient to warrant further investigation and, with appropriate supervision, a useful aid to clinical work in its current state of development. It is not yet of a standard to warrant claims of evidence of treatment progress in itself.

This study investigated the position and movement of individuals with respect to their clinical engagement in a group-based program, using the construct of DMS as an indicator. Therapist style factors might play a part in any such movement, but this remains to be investigated. Also, the impact of such shifts on group dynamics was outside of the scope of this article. From the general literature surrounding the characteristics of therapy groups, it is accepted that a shift in the behavior of a subsection of the group can have implications for the progress of others. The overall positive shift in DMS observed in this study might be partly explained by such group dynamics. The potential for maximizing treatment progress by more conscious attention to group dynamics might contribute to more efficient and expedient yield in offender work. In the interests of efficiency at least then, investigation of group relational factors in promoting an exploratory climate in the group is worthy of consideration.

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.

References

Allan

Grace

R. C.

Rutherford

Hudson

S. M.

(2007). Psychometric assessment of dynamic risk factors for child molesters. Sexual Abuse: A Journal of Research and Treatment, 19, 347-367.

Andrews

Bonta

(2010). The psychology of criminal conduct (5th ed.). Cincinnati, OH: Anderson.

Barrett

Wilson

R. J.

Long

(2003). Measuring motivation to change in sexual offenders from institutional intake to community treatment. Sexual Abuse: A Journal of Research and Treatment, 15, 269-283.

Beech

A. R.

Hamilton-Giachritsis

C. E.

(2005). Relationship between therapeutic climate and treatment outcome in group-based sexual offender treatment programs. Sexual Abuse: A Journal of Research and Treatment, 17, 127-139.

Beggs

S. M.

(2010). Within-treatment outcome among sexual offenders: A review. Aggression and Violent Behavior, 15, 369-379.

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.

Burlingame

Strauss

Joyce

(2013). Change mechanisms and effectiveness of small group treatment. In Lambert

M. J.

(Ed.), Bergin and Garfield’s handbook of psychotherapy and behavior change (7th ed., pp. 640-689). Hoboken, NJ: John Wiley.

Craissati

Beech

A. R.

(2003). A review of dynamic variables and their relationship to risk prediction in sex offenders. Sexual Aggression, 9, 41-45.

Drieschner

K. H.

Verschuur

(2010). Treatment engagement as a predictor of premature treatment termination and treatment outcome in a correctional outpatient sample. Criminal Behaviour and Mental Health, 20, 86-99.

10.

Frost

(2004). Therapeutic engagement styles of child sexual offenders in a group treatment program: A grounded theory study. Sexual Abuse: A Journal of Research and Treatment, 16, 191-208.

11.

Frost

Connolly

(2004). Reflexivity, reflection, and the change process in offender work. Sexual Abuse: A Journal of Research and Treatment, 16, 365-380.

12.

Frost

Daniels

Hudson

S. M.

(2006). Disclosure strategies among sex offenders: A model for understanding the engagement process in groupwork. Journal of Sexual Aggression, 12, 227-244.

13.

Grady

M. D.

Brodersen

Abramson

J. M.

(2011). The state of psychological measures for adult sexual offenders. Aggression and Violent Behavior, 16, 227-240.

14.

Hanson

R. K.

Thornton

(2000). Improving risk assessments for sex offenders: A comparison of three actuarial scales. Law and Human Behavior, 24, 119-136.

15.

Harkins

Beech

A. R.

(2007). A review of the factors that can influence the effectiveness of sexual offender treatment: Risk, need, responsivity, and process issues. Aggression and Violent Behavior, 12, 615-627.

16.

Helmus

Thornton

Hanson

Babchishin

(2012). Improving the predictive accuracy of Static-99 and Static-2002 with older sex offenders: Revised age weights. Sexual Abuse: A Journal of Research and Treatment, 24, 64-101.

17.

Holdsworth

Bowen

Brown

Howat

(2014). Offender engagement in group programs and associations with offender characteristics and treatment factors: A review. Aggression and Violent Behavior, 19, 102-121.

18.

Hudson

S. M.

Ward

T. L.

Marshall

W. W.

Johnston

Jones

(1995). Kia Marama: A Cognitive-Behavioural Program for Incarcerated Child Molesters. Behaviour Change, 12(2), 69-80.

19.

Hudson

S. M.

Wales

D. S.

Ward

(1998). Kia Marama: A treatment program for child molesters in New Zealand. In Marshall

W. L.

Fernandez

Y. M.

Hudson

S. M.

Ward

(Eds.), Sourcebook of treatment for sexual offenders (pp. 17-28). New York, NY: Plenum.

20.

Kim

Benekos

Merlo

(2016). Sex offender recidivism revisited: Review of recent meta-analyses on the effects of sex offender treatment. Trauma, Violence, & Abuse, 17, 105-117.

21.

Marshall

W. L.

Burton

D. L.

(2010). The importance of group processes in offender treatment. Aggression and Violent Behavior, 15, 141-149.

22.

McGrath

R. J.

Cumming

G. F.

Burchard

B. L.

Zeoli

Ellerby

(2010). Current practices and emerging trends in sexual abuser management: The safer society 2009 North American Survey. Brandon, VT: The Safer Society Press.

23.

McGrath

R. J.

Hoke

S. E.

Vojtisek

S. E.

(1998). Cognitive-behavioral treatment of sex offenders: A treatment comparison and long-term follow-up study. Criminal Justice and Behavior, 25, 203-225.

24.

McMurran

Ward

(2010). Treatment readiness, treatment engagement and behaviour change. Criminal Behaviour and Mental Health, 20, 75-85.

25.

Olver

M. E.

Wong

S. C. P.

Nicholaichuk

T. P.

Gordon

A. E.

(2007). The validity and reliability of the Violence Risk Scale–Sexual Offender version: Assessing sex offender risk and evaluating therapeutic change. Psychological Assessment, 19, 318-329.

26.

Polaschek

D. D.

(2012). An appraisal of the risk–need–responsivity (RNR) model of offender rehabilitation and its application in correctional treatment. Legal and Criminological Psychology, 17, 1-17.

27.

Reeves

Ogloff

Simmons

(2017). The predictive validity of the Static-99, Static-99R, and Static-2002/R: Which one to use? Sexual Abuse: A Journal of Research and Treatment, 30, 887-907. doi:10.1177/1079063217712216

28.

Scott

K. L.

King

C. B.

(2007). Resistance, reluctance, and readiness in perpetrators of abuse against women and children. Trauma, Violence, & Abuse, 8, 401-417.

29.

Serran

G. A.

Marshall

W. M.

(2010). Therapeutic process in the treatment of sexual offenders: A review article. The British Journal of Forensic Practice, 12(3), 4-16.

30.

Skelton

Riley

Wales

Vess

(2006). Assessing risk for sexual offenders in New Zealand: Development and validation of a computer-scored risk measure. Journal of Sexual Aggression, 12, 277-286.

31.

Storey

Watt

Jackson

Hart

(2012). Utilization and implications of the Static-99 in practice. Sexual Abuse: A Journal of Research and Treatment, 24, 289-302.

32.

Walji

Simpson

Weatherhead

(2013). Experiences of engaging in psychotherapeutic interventions for sexual offending behaviors: A meta-synthesis. Journal of Sexual Aggression, 20, 310-332.

33.

Ware

(2011). The importance of contextual issues within sexual offender treatment. In Boer

D. P.

Eher

Craig

L. A.

Milner

M. H.

Pfäfflin

(Eds.), International perspectives on the assessment and treatment of sexual offenders: Theory, practice and research (pp. 299-312). Chichester, UK: John Wiley.

34.

Ware

Mann

R.E.

(2012). How should “acceptance of responsibility” be addressed in sexual offending treatment programs? Aggression and Violent Behavior, 17(4), 279-288.

Therapist	N	Openness	Self-directedness
A	8	.87	.74
B	7	.95	.96
C	54	.62	.42
D	55	.77	.86
E	48	.62	.40
F	24	.44	.78
G	28	.77	.78
All	224	.66	.70

Therapist	N	Openness	Self-directedness
A	8	.87	.74
B	7	.95	.96
C	54	.62	.42
D	55	.77	.86
E	48	.62	.40
F	24	.44	.78
G	28	.77	.78
All	224	.66	.70

Therapist	N	Openness	Self-directedness
A	8	.87	.74
B	7	.95	.96
C	54	.62	.42
D	55	.77	.86
E	48	.62	.40
F	24	.44	.78
G	28	.77	.78
All	224	.66	.70