Dosage of treatment to sexual offenders: Are we overprescribing?

International Journal of Offender Therapy and Comparative Criminology http://ijo.sagepub.com/ Dosage of Treatment to Sexual Offenders: Are We Overprescribing? Donna L. Mailloux, Jeffrey Abracen, Ralph Serin, C. Cousineau, Bruce Malcolm and Jan Looman Int J Offender Ther Comp Criminol 2003 47: 171 DOI: 10.1177/0306624X03251096 The online version of this article can be found at: http://ijo.sagepub.com/content/47/2/171 Published by: http://www.sagepublications.com Additional services and information for International Journal of Offender Therapy and Comparative Criminology can be found at: Email Alerts: http://ijo.sagepub.com/cgi/alerts Subscriptions: http://ijo.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://ijo.sagepub.com/content/47/2/171.refs.html Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 Dosage of Treatment to Sexual Offenders: Are We Overprescribing? Donna L. Mailloux Jeffrey Abracen Ralph Serin C. Cousineau Bruce Malcolm Jan Looman Abstract: A sample of 337 offenders who received treatment in a variety of sex offender treatment programs in the Ontario region of Correctional Service Canada between 1993 and 1998 were divided based on the highest intensity sex offender programming that they received (low, moderate, and high). The three groups were compared with reference to a variety of actuarial risk assessment measures, criminogenic factors, and the number and type of treatment programs completed. It was hypothesized that the high-intensity group would have more criminogenic risk factors, higher actuarial scores, and participate in more treatment programs than both the moderate- and low-intensity groups. The results indicate that in general, the hypotheses were supported. Nonetheless, the results suggest that the low-intensity group may be receiving too much sex offender-specific treatment. Keywords: sex offenses; treatment Within the past number of years several reviews of the sex offender treatment literature have been published (Abracen & Looman, in press; Alexander, 1999; Hall, 1995; Hanson, 2000; Marshall, Anderson, & Fernandez, 1999; Quinsey, Harris, Rice, & Cormier, 1998). With the exception of Quinsey et al. (1998), these reviews have concluded that the risk of recidivism is reduced with the application of contemporary cognitive-behavioral treatment techniques. Nevertheless, the long-standing debate regarding treatment efficacy in general has hindered the investigation of other important treatment issues. The consistency of the outcome studies with sex offenders accentuates the need to move beyond basic questions regarding sex offender treatment (Abracen & Looman, in press) in an effort to inform a strategy for the provision of treatment to sexual offenders. Several authors have already begun to investigate factors associated with sex offender treatment in an effort to improve on existing protocols. For example, NOTE: Please address correspondence to Jeffrey Abracen, Ph.D., Central District Parole, Department of Psychology, 330 Keele St., Toronto, Ontario, Canada M6P 2K7; e-mail: [email protected]. International Journal of Offender Therapy and Comparative Criminology, 47(2), 2003 DOI: 10.1177/0306624X03251096  2003 Sage Publications 171-184 171 Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 172 International Journal of Offender Therapy and Comparative Criminology with high-risk clients, longer treatment programs may be more effective (Lösel, 1998). As well, individuals with high Hare Psychopathy Checklist-Revised (PCL-R) scores (e.g., above 25) may be best served with inpatient treatment (Hare, 1998). Additional work relevant to these topics and other related issues are important if treatment providers are to improve the quality of their programs. Mailloux and Serin (2001) reported one example of research related to important clinical questions relevant to the treatment of sexual offenders. This study investigated the determinants of sex offender treatment participation. The sample consisted of sex offenders incarcerated in the Ontario region of Correctional Service of Canada (CSC), which provides services to offenders serving sentences of over 2 years. Mailloux and Serin found that 67% of sex offenders take sex offender programs with a completion rate of 86.9%. On average, sex offenders took 3.2 different programs and 6.7 programs overall, reflecting some repetition. Whereas those of higher risk (based on actuarial assessment) participated in a greater variety and quantity of programs, deniers and rapists were significantly less likely to take sex offender programs. Furthermore, although sex offender program participation differentiated recidivists from nonrecidivists, it was not predictive of outcome when controlling for risk and index sex offence (rapist and intra- and extrafamilial offender). It is interesting to note that among offenders who did not take sex offender treatment, when other programs were taken their outcome was worse than if no programs were taken. These results raise some important issues regarding the allocation of sex offender treatment programs. First, although Mailloux and Serin (2001) found that higher risk offenders attended a greater number of treatment programs (including more sex offender treatment programs), rapists were less likely to receive treatment. This raises the possibility that programming is allocated differentially to particular groups of higher risk offenders. Second, a related question involves whether the allocation of sex offender treatment is appropriately matched to the offenders’ particular risk level (e.g., low, moderate, or high). This is a salient issue given that many correctional systems provide programs at various intensity levels. In an effort to understand the dynamics associated with the findings of Mailloux and Serin (2001), the relationship between dosage of treatment, risk, and outcome need to be investigated. For instance, if treatment prescription is consistent with the principles of risk and need, then those of higher risk/need should be in higher intensity programs (see Andrews & Bonta, 1998, for a discussion). Conversely, lower risk offenders should attend lower intensity treatment programs. In the context of the current investigation, dosage refers to the number of treatment programs completed or the intensity of the intervention or both. Within the Ontario region of CSC, sex offenders admitted into federal correctional facilities are assessed at the Millhaven Assessment Unit (MAU). The purpose of the assessment is to determine an offender’s institutional placement and to prescribe an individualized treatment plan based on the principles of risk/need/ responsivity. Treatment should be matched to the risk level and learning style of Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 Dosage of Treatment 173 the offender while ensuring that criminogenic needs are being targeted. Moreover, the assessment includes a recommendation for a particular intensity of sex offender program (low, moderate, or high). Within CSC, specific criteria exist for programs at each intensity level. Programs at higher intensity levels require more hours of treatment and thus extend for significantly greater periods of time. At the time the data in the present study were collected, the range for program duration was from approximately 2 months (for low-intensity programs) to between 6 and 8 months (for high-intensity programs). The Ontario region of CSC is perhaps the best jurisdiction to conduct such an investigation given the range of outcome data available. For example, all individuals in the current study identified as attending high-intensity programs attended one of two programs offered at the Regional Treatment Centre (RTCSOTP). The RTCSOTP offers an inpatient-based treatment program consisting of both individual and group therapy. For clients who are not able to attend the full treatment program (e.g., due to preexisting psychiatric conditions), an individual therapy only program is available. Two outcome studies have been conducted regarding the efficacy of these programs (Di Fazio, Abracen, & Looman, 2001; Looman, Abracen, & Nicholaichuk, 2000). For example, Looman et al. (2000) demonstrated that relative to matched untreated comparison participants, individuals attending the RTCSOTP (both individual and group treatment) recidivated at less than half the rate of comparison participants. These data were interpreted to be both statistically and clinically significant. Follow-up data were presented for approximately a 10-year period. With reference to individuals attending moderate-intensity programs, in the current investigation the vast majority of participants completed either the Bath Institution Sex Offender Treatment Program and/or the Warkworth Sexual Behavior Clinic Treatment Program (WSBC). Both of these programs have been discussed extensively in the literature. With reference to the WSBC program, three outcome studies are available (Barbaree, Seto, & Langton, 2001; Barbaree, Seto, Langton, & Peacock, 2001; Seto & Barbaree, 1999). Dr. W. L. Marshall has written extensively about the effectiveness of the program offered at Bath Institution (e.g., see discussion in Marshall et al., 1999). Although controlled outcome data are not available regarding recidivism with this population, other sex offender programs offered by Marshall and his colleagues have received empirical support (e.g., Marshall & Barbaree, 1988; Marshall, Eccles, & Barbaree, 1991). Furthermore, pre- and post-treatment psychometric data presented by Marshall and his colleagues regarding the Bath Institution Sex Offender Program have been encouraging (e.g., Marshall, Bryce, Hudson, Ward, & Moth, 1996). Although no outcome data exist regarding participants attending the lowintensity sex offender programs identified in the current study, there is sound theoretical reason to believe that such outcome data might not yield statistically significant results even if the programs were clinically efficacious. The difficulty with the evaluation of low-intensity programs is that given the low base rate of sexual offending in general (see Hanson & Bussière, 1998), the treatment effect Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 174 International Journal of Offender Therapy and Comparative Criminology necessary to demonstrate statistical significance must be unreasonably large or follow-up must be conducted for excessive periods of time. These issues are compounded when participants who are at low risk of recidivism (e.g., incest offenders) are overrepresented in such samples. Barbaree (1997) provided a detailed discussion of the statistical complexities associated with demonstrating significance in populations that have very low base rates. The reader is referred to this work for a more detailed discussion of these matters. The Ontario region of CSC represents a jurisdiction that has produced more outcome data regarding sex offender treatment than perhaps any other jurisdiction that exists today. As such, the Ontario region of CSC represents an ideal environment in which to investigate whether a system of programs, many of which have received empirical support, is allocated in a theoretically meaningful way. The following study may be viewed as an attempt to investigate resource allocation assessment using measures that have been empirically linked with outcome. The present study investigates the allocation of sex offender treatment to a large group of sexual offenders. As noted earlier, sex offenders attending highintensity sex offender programs should be assessed as being higher risk and as having more criminogenic needs than sex offenders in lower intensity sex offender programs. Therefore, the following hypotheses were investigated: Hypothesis 1: Sex offenders attending high-intensity sex offender treatment programs would have more static criminogenic risk factors than sex offenders attending moderate-or low-intensity sex offender treatment programs. Hypothesis 2: Sex offenders in higher intensity sex offender programs would have higher scores on actuarially based risk assessments than sex offenders attending moderate- or low-intensity sex offender treatment programs. Hypothesis 3: Sex offenders attending high-intensity sex offender treatment programs would participate in more sex offender-specific programs and more programs related to other criminogenic needs than offenders in low- and moderate-intensity sex offender treatment programs. METHOD PARTICIPANTS Participants consisted of all offenders assessed at MAU between 1993 and 1998. By definition, all participants were incarcerated in Canadian federal penitentiaries. The initial sample consisted of 502 consecutive admissions. For the purpose of this investigation, a subset of 354 offenders was selected based on their participation (vs. completion) in at least one sex offender treatment program (see the following for a discussion). Program intensity designation was based on two criteria: (a) discussions with managers of sex offender treatment programs at both the national and regional Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 Dosage of Treatment 175 (i.e., the Ontario region of CSC) levels and (b) CSC standards (National Committee on the Provisions of Sex Offender Treatment, 1996) that were in place at the time of the current investigation. Intensity groups were then created based on the highest level of intensity of sex offender treatment completed during the study period. For example, those who participated in both a low-intensity and a moderateintensity treatment program were classified as moderate intensity. Sex offenders who had received primarily individualized treatment in a program with unspecified content and no group treatment were excluded from the analyses (n = 17). The following intensity groups were subsequently created: low intensity (n = 38), moderate intensity (n = 265), and high intensity (n = 34). PROCEDURES The files of all sex offenders assessed at MAU between 1993 and 1998 were coded for information pertaining to background and criminal history. Risk assessment scores on the Level of Service Inventory (LSI) (Bonta & Andrews, 1993) and the Psychopathy Checklist-Revised (Hare, 1991) were taken from the MAU files. These assessments were conducted as part of the MAU assessment and were based on both an interview and extensive file review. Programming information was obtained from the Offender Management System (OMS). OMS is an electronic database containing information related to all offenders under the jurisdiction of CSC. Programming information was grouped according to the type of correctional program (i.e., cognitive skills, substance abuse, and sexual offender). RESULTS CRIMINOGENIC FACTORS The purpose of this section is to determine whether the intensity groups differed on meaningful criminogenic factors. These factors include the following static risk factors: age at index offense, number of nonsexual violent convictions, number of nonviolent convictions, number of sexual convictions, marital status, type of index sex offense (rapist, and extrafamilial and intrafamilial child molester), early onset of sexual offending, juvenile record, and convictions for diverse sex crimes. A 3 × 4 multivariate analysis of variance (MANOVA) was conducted with the intensity group (low, moderate, and high) as the independent variable and age at index offense, number of nonsexual violent offenses, number of nonviolent offenses, and number of sexual offenses as the dependent variables. All the dependent variables were moderately correlated, justifying the use of the MANOVA with these variables. The overall MANOVA was significant, Wilks F(8, 594) = 5.75, p < .01; η2 = .07, suggesting a significant but low association between the dependent and independent variables. An examination of the Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 176 International Journal of Offender Therapy and Comparative Criminology univariate tests revealed significant main effects across all variables except number of sexual convictions (Table 1). A post hoc investigation using Tamhane’s test (given that the assumption of homogeneity of the variance-covariance matrices could not be guaranteed) supported the hypothesis that the low-intensity group had significantly fewer nonviolent convictions than both the moderate- (mean difference = –7.30, p < .01) and high-intensity (mean difference = –13.11, p < .01) groups. No differences were found between the moderate- and high-intensity groups. A similar pattern of results was found for both nonsexual violent convictions and age at index offense. The low-intensity group had significantly fewer nonsexual violent convictions than both the moderate- (mean difference = –1.41, p < .01) and high-intensity groups (mean difference = –1.98, p < .01). No significant difference was observed between the moderate- and high-intensity groups. Moreover, the low-intensity group was found to be significantly older than both the moderate- (mean difference = 5.31, p < .01) and high-intensity (mean difference = 5.80, p < .05) groups, but again no differences emerged between the moderate- and high-intensity groups. A series of chi-square analyses were performed on the categorically coded data. The first 3 × 3 chi-square analysis was performed to determine whether the intensity groups varied with reference to marital status (married/common law, separated/widowed/divorced, and single). The overall test was significant, χ2(4) = 13.14, p < .05. An examination of the standardized residuals revealed that lowintensity offenders were significantly less likely to be single and significantly more likely to be married or in common-law relationships. The second 3 × 3 chi-square analysis was computed to determine whether the three intensity level groups differed with reference to index offence category (rapists, extrafamilial, and intrafamilial). The chi-square was significant, χ2(4) = 33.16, p < .001. An examination of the standardized residuals indicated that there were proportionally more intrafamilial child molesters and proportionally fewer rapists in the low-intensity group than would be expected by chance. Moreover, the high-intensity group consisted of proportionally fewer intrafamilial child molesters than would be expected by chance. Referring to Table 2, a series of 3 × 2 chi-squares were performed to determine whether the intensity groups differed in their history of early onset of sexual offending (younger than 19, yes/no), presence of a juvenile record (yes/no), and convictions for diverse sex crimes (yes/no). The chi-square analyses were significant, demonstrating the following pattern of results: Low-intensity offenders were proportionally less likely to have a juvenile record, whereas the highintensity group was proportionally more likely to have an early onset of sexual offending. None of the other proportional differences reached acceptable levels based on examination of the standardized residuals. Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 Dosage of Treatment 177 TABLE 1 CRIMINOGENIC FACTORS Moderate Intensity (n = 240) Low Intensity (n = 34) Age at index offencea Number nonsexual violent offencesb Number nonviolent offencesc Number of sexual offences High Intensity (n = 29) M SD M SD M SD 36.18 9.19 30.87 8.52 30.38 9.21 0.13 1.00 3.74 0.37 2.32 2.92 1.54 8.29 4.31 2.49 10.34 6.19 2.11 14.10 5.45 2.95 14.13 6.14 2 a. Univariate main effect, F(2, 300) = 5.82, p < .01, η = .04. 2 b. Univariate main effect, F(2, 300) = 6.36, p < .01, η = .04. 2 c. Univariate main effect, F(2, 300) = 13.244, p < .01, η = .08. TABLE 2 COUNTS (PERCENTAGE BY INTENSITY LEVEL IN PARENTHESES) AND STANDARDIZED RESIDUALS (SR) BY GROUP: DICHOTOMOUS DATA Juvenile historya Yes No Early onset of sexual offendingb Yes No Diverse sex crimesc Yes No Low Intensity Moderate Intensity High Intensity 4 (10.5) SR: –2.4 34 (89.5) SR: 1.9 90 (34.6) SR: 0.4 166 (63.8) SR: –0.2 16 (50) SR: 1.6 14 (43.8) SR: –1.5 0 (0) SR: –1.9 38 (100) SR: 0.7 25 (9.6) SR: 0 233 (89.6) SR: 0.0 7 (21.9) SR: 2.2 24 (75) SR: –0.9 2 (5.3) SR: –1.2 36 (94.7) SR: 0.6 32 (12.3) SR: 0.1 225 (86.5) SR: 0.1 6 (18.8) SR: 1.1 22 (68.8) SR: –1.0 2 a. χ (4) = 18.78, p < .01. 2 b. χ (4) = 11.93, p < .05. 2 c. χ (4) = 22.30, p < .01. Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 178 International Journal of Offender Therapy and Comparative Criminology RISK MEASURES To address the hypothesis that the intensity groups would differ on actuarial measures of risk, two separate analyses of variance (ANOVAs) were performed using intensity as the independent variable and the LSI and PCL-R scores as dependent measures. ANOVAs were performed instead of a MANOVA given the high correlation (r = .83) between the dependent measures. For highly correlated data such as these, Tabachnick and Fiddell (2001) recommend the use of univariate tests. The overall ANOVA was significant for both the PCL-R, F(2, 226) = 18.25, p < .001, η2 = .14, and the LSI, F(2, 235) = 19.07, p < .001, η2 = .14. Referring to Table 3, post hoc testing (Tamhane) indicated that as hypothesized, the high-intensity group scored significantly higher on the PCL-R than both the moderate- (mean difference = 4.93, p < .01) and low-intensity (mean difference = 11.18, p < .001) groups. The moderate-intensity group also scored significantly higher on the PCL-R than the low-intensity group (mean difference = 6.25, p < .001). Post hoc testing (Tamhane) on the LSI demonstrated a similar pattern of results, with the high-intensity group scoring significantly higher than both the moderate- (mean difference = 5.09, p < .05) and low-intensity (mean difference = 13.54, p < .001) groups. The moderate-intensity group also scored significantly higher on the LSI than the low-intensity group (mean difference = 8.45, p < .001). CORRECTIONAL PROGRAMMING To address the hypothesis that the intensity groups would differ on the number of correctional treatment programs completed, data were analyzed for both the number and types of programs taken. Data were only analyzed for programs that related specifically to criminogenic needs (i.e., substance abuse, sex offenderspecific program, and cognitive skills) and that were attended by more than half the sample. This programming information was available for only 253 offenders in the sample. Means and standard deviations are listed in Table 4. A 3 × 3 between-subjects MANOVA was performed with intensity level (high, moderate, and low) as the independent variable and number of programs (substance abuse, sex offender-specific program, and cognitive skills) completed as the dependent variable. With the use of the Wilks criterion, the dependent variables were significantly different by intensity level, F(6, 496) = 4.93, p < .001. A weak effect size was observed, η2 = .06. Univariate tests indicated that there were significant differences between the groups with reference to the number of substance abuse programs, F(2, 250) = 6.7, p < .01; sex offender programs, F(2, 250) = 4.31, p < .05; and cognitive skills programs completed, F(2, 250) = 5.0, p < .01. Post hoc comparisons using Tukey’s honestly significant difference demonstrated significant differences between the intensity levels within the three types of programs. For substance abuse programs, the low-intensity group took significantly fewer programs than either moderate- (mean difference = –.44, p < .01) or high-intensity (mean difference = –.53, p < .05) groups. No significant differences were found Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 Dosage of Treatment 179 TABLE 3 MEANS AND STANDARD DEVIATIONS FOR ACTUARIAL RISK ASSESSMENT SCORES Low Intensity Psychopathy ChecklistRevised total Level of Service Inventory total Moderate Intensity High Intensity M SD n M SD n M SD n 11.17 4.53 35 17.42 7.52 174 22.35 5.58 20 13.03 6.04 35 21.48 9.25 182 26.57 7.41 21 TABLE 4 MEANS AND STANDARD DEVIATIONS FOR CORRECTIONAL PROGRAMS BY INTENSITY LEVEL Low Intensity (n = 37) Substance abuse Cognitive skills Sex offender-specific program Moderate Intensity (n = 195) High Intensity (n = 21) M SD M SD M SD 0.32 1.16 1.51 0.63 1.24 0.65 0.77 1.96 1.49 0.70 1.42 0.70 0.86 1.95 1.95 0.85 1.66 0.59 between the moderate- and high-intensity groups. For sex offender programs, both the low-intensity (mean difference = –.44, p < .05) and moderate-intensity group (mean difference = –.46, p < .05) took significantly fewer programs than the high-intensity group. There were no significant differences between the low- and moderate-intensity groups. For the cognitive skills programs, the low-intensity group took significantly fewer programs than the moderate-intensity group (mean difference = –.80, p < .01). No other groups differed from each other on cognitive skills programs. DISCUSSION The results of the present investigation suggest that dosage may be a relevant issue to consider with reference to program allocation. By adopting a system such as that described earlier, important issues related both to responsivity (Andrews & Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 180 International Journal of Offender Therapy and Comparative Criminology Bonta, 1998) and cost-effectiveness might be addressed. For example, given that there are significantly greater costs associated with the inpatient-based highintensity program offered at the RTCSOTP versus low-intensity sex offender programming offered in a minimum-security institution, the savings associated with the appropriate allocation of treatment would certainly offset the costs associated with the system of triage described earlier in any large correctional jurisdiction. The results of the present study provide support for the stated hypotheses. Significant differences between the three intensity groups were observed with reference to both the PCL-R and the LSI. The high-intensity group evidenced significantly higher scores than either of the other groups on both of these measures. The low-intensity group evidenced significantly lower scores than either of the other groups on these measures as well. That the high-intensity group evidenced mean PCL-R scores that bordered on the recommended cutoff of 25 for psychopathy (Quinsey et al., 1998) is of relevance as well. This suggests that individuals who are at a particularly high risk for recidivism are receiving more intense treatment programs. In general, the data suggest a significant degree of overlap between the moderate- and high-intensity groups on a number of the variables investigated in the present study. For example, these groups did not differ with reference to number of nonviolent or nonsexual violent offences. Contrary to prediction, the three intensity groups did not differ significantly with reference to number of sexual convictions. Nonetheless, the group means were in the expected direction, with higher intensity participants having more sexual convictions on their records. Furthermore, moderate- and high-intensity offenders were proportionally similar with reference to marital history. The low-risk group was, however, proportionally more likely to be married. Given that being in an intimate relationship appears to be a protective factor (Marshall et al., 1999), this finding is in keeping with the stated hypotheses of the study (i.e., that low-intensity clients would present a lower risk). As well, no significant differences were observed between the moderate- and high-intensity groups with reference to either the number of cognitive skills or substance abuse programs taken. The fact that there is overlap between the moderate- and high-intensity groups is not surprising given that the distinctions drawn between individuals assigned to the various intensity-level programs are to an extent based on arbitrary criteria derived from both the empirical literature and clinical judgment. The fact that the clearest distinctions between the three groups were with reference to actuarially based risk assessment measures suggests that this is one of the primary means by which sex offenders are assigned to the various intensity levels. This strategy is clearly defensible from a theoretical perspective in that higher intensity programming should be reserved for higher risk clients (Andrews & Bonta, 1998). Furthermore, criminal history data are only one indicator of risk. Both of the risk assessment measures used in the current investigation considered criminal history data and background factors believed to be associated with risk as well. As such, Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 Dosage of Treatment 181 they provide a more inclusive assessment of risk than one based on criminal history alone. In terms of correctional programming, the number of sex offender treatment programs taken is relevant. Results indicated that the high-intensity group received more sex offender programming than either the low- or moderate-intensity groups. No significant differences were observed between the low- and moderateintensity groups in terms of sex offender programming. One interpretation of these data is that low-intensity participants are receiving too much treatment relative to the other intensity levels. In fact, with reference to sex offender-specific treatment, the group mean for low-intensity participants was marginally higher than for moderate-intensity participants. Furthermore, the low-intensity group received fewer cognitive skills programs relative to the moderate group but did not differ significantly from the high-intensity group on this variable. In hindsight, it is perhaps not surprising to find overlap between moderate- and high-intensity offenders regarding correctional programming. Both of these groups presented with significant criminal histories and were composed of sex offenders belonging to higher risk groups (i.e., rapist and extrafamilial child molesters). Also, low-intensity offenders seemed to differ from both the moderate- and highintensity offenders on a number of important background variables related to risk (e.g., age at incarceration and number of nonsexual violent offences). Furthermore, the low-intensity group scored significantly lower on actuarial measures of risk than either of the other groups. In spite of these important differences, the low-intensity group received a similar number of sex offender treatment programs relative to the moderate-intensity group and a similar number of cognitive skills programs relative to the high-intensity group. These data imply that we may be overprescribing treatment to low-intensity offenders. It should also be noted that intrafamilial sex offenders were overrepresented in the low-intensity group. The fact that intrafamilial offenders have very low recidivism rates relative to other groups of sex offenders (e.g., Hanson & Bussière, 1998) underscores the aforementioned observation. The data from this study alone do not justify the conclusion that low-intensity offenders may be receiving too many sex offender treatment programs relative to the other intensity levels. However, the research cited earlier regarding the efficacy of programs included in both the moderate- and high-intensity groups clearly adds credibility to this hypothesis. Although the authors are not aware of any outcome data related to the low-intensity sex offender programs included in the present analyses, the fact that it would be very difficult in practice to establish any statistically significant differences between individuals assigned to the low-intensity programs only and a comparison group is in itself relevant to this conclusion (see Barbaree, 1997, for a discussion of related matters). The reader should keep in mind that the methodology used in the current investigation ensured that individuals assigned to the low-intensity group had not received any sex offender treatment programs labeled as either moderate or high intensity. Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 182 International Journal of Offender Therapy and Comparative Criminology Based on the available data it would seem that both moderate- and highintensity clients have significant criminal histories. Furthermore, the mean PCLR score for high-intensity offenders borders on a recommended cutoff for psychopathy. Moderate- and high-intensity offenders’ scores on both the LSI and PCL-R, as well as their criminal histories, imply that these individuals present with significant treatment needs. It seems reasonable to conclude that many clients at these intensity levels require more treatment. In short, the data suggest high-intensity offenders should receive additional treatment relative to lower risk offenders. This may mean more programs as well as programs of longer duration. Sequencing of programs may also be a consideration, particularly for higher risk offenders. For example, it may be useful for offenders to complete a program such as cognitive skills (which introduces basic elements associated with cognitivebehavioral therapy as well as concrete suggestions as to how to apply these principles to everyday situations) prior to completing more intensive therapeutic programs. The present study may be viewed as an investigation of resource allocation among sex offenders. The data included in this investigation may be relevant for other jurisdictions that are interested in developing an empirically based system of treatment for sexual offenders. As well, the data included earlier might be used as guidelines for cutoffs for some of the best known risk assessment instruments available today (i.e., the PCL-R and LSI). The current investigation contributes to the literature regarding treatment efficacy with sexual offenders by investigating an issue that has not received a great deal of attention in the literature, namely, dosage of treatment. The possibility that certain groups of sexual offenders may require more than one sex offender treatment program whereas others may be receiving too many programs relative to their risk level was addressed in the current investigation. Given the paucity of research regarding dosage of treatment to sexual offenders, the current investigation provides a framework based on empirically validated indices that others may use to guide treatment decisions regarding sexual offenders. By comparing individuals who attend treatment programs with demonstrated efficacy but that differ with reference to intensity, the current investigation contributes to research related to the treatment of sexual offenders that moves beyond the basic question as to whether treatment works. Nonetheless, this study represents preliminary work in this area, and many related issues need to be addressed. For example, whether pre- and posttreatment psychometric testing can reliably distinguish treatment successes from failures (e.g., with reference to recidivism) needs further attention. Clinicians may be reluctant to treat offenders who present with the highest risk or need levels such as those who attended the high-intensity program described in this study. As well, clients presenting with the lowest risk or need levels may still present with very real treatment issues. Nonetheless, if the primary goal of forensic treatment is to reduce recidivism, then the question becomes one regarding the appropriate allocation of limited resources. Given this real-world concern, atten- Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 Dosage of Treatment 183 tion to criminogenic factors (e.g., those identified by the PCL-R and LSI) may be the most rational basis on which to determine treatment need at present. That is, resourcing should be based on the identification of risk for recidivism and number or intensity of criminogenic factors. REFERENCES Abracen, J., & Looman, J. (in press). Issues in the treatment of sexual offenders: Recent developments and directions for future research. Aggression and Violent Behavior. Alexander, M. A. (1999). Sexual offender treatment efficacy revisited. Sexual Abuse: A Journal of Research and Treatment, 11, 101-116. Andrews, D. A., & Bonta, J. (1998). The psychology of criminal conduct (2nd ed.). Cincinnati, OH: Anderson. Barbaree, H. (1997). Evaluating treatment efficacy with sexual offenders: The insensitivity of recidivism studies to treatment effects. Sexual Abuse: A Journal of Research and Treatment, 9, 111-128. Barbaree, H. E., Seto, M. C., & Langton, C. M. (2001, October). Psychopathy, treatment behavior and sex offender recidivism: Extended follow-up. Paper presented at the annual meeting of the Association for the Treatment of Sexual Abusers, San Antonio, TX. Barbaree, H. E., Seto, M. C., Langton, C. M., & Peacock, E. J. (2001). Evaluating the predictive accuracy of six risk assessment instruments for adult sex offenders. Criminal Justice and Behavior, 28, 490-521. Bonta, J., & Andrews, D. A. (1993). The level of supervision inventory: An overview. IARCA Journal, 5, 6-8. Di Fazio, R., Abracen, J., & Looman, J. (2001). Group versus individual treatment of sex offenders: A comparison. Forum on Corrections Research, 13, 56-59. Hall, G. C. N. (1995). Sexual offender recidivism revisited: A meta-analysis of recent treatment studies. Journal of Consulting and Clinical Psychology, 63, 802-809. Hanson, R. K. (2000, November). The effectiveness of treatment for sexual offenders: Report of the ATSA collaborative data research committee. Paper presented at the annual meeting of the Association for the Treatment of Sexual Abusers, San Diego, CA. Hanson, R. K., & Bussière, M. T. (1998). Predicting relapse: A meta-analysis of sexual offender recidivism studies. Journal of Consulting and Clinical Psychology, 66, 348-362. Hare, R. D. (1998). Psychopaths and their nature: Implications for the mental health and criminal justice systems. In T. Millon, E. Simonsen, M. Birket-Smith, & R. D. Davis (Eds.), Psychopathy: Antisocial, criminal and violent behavior (pp. 188-212). New York: Guilford. Hare, R. D. (1991). The Hare Psychopathy Checklist-Revised. Toronto, Canada: Multi-Health Systems. Looman, J., Abracen, J., & Nicholaichuk, T. P. (2000). Recidivism among treated sexual offenders and matched controls: Data from the Regional Treatment Centre (Ontario). Journal of Interpersonal Violence, 15, 279-290. Lösel, F. (1998). Treatment and management of psychopaths. In D. J. Cooke, A. E. Forth, & R. D. Hare (Eds.), Psychopathy: Theory research and implications for society (pp. 303-354). London: Kluwer Academic Publishers. Mailloux, D. L., & Serin, R. C. (2001). Issues impacting on programming for sex offenders. Manuscript submitted for publication. Marshall, W. L., Anderson, D., & Fernandez, Y. (1999). Cognitive behavioral treatment of sexual offenders. New York: John Wiley. Marshall, W. L., & Barbaree, H. E. (1988). The long-term evaluation of a behavioral treatment program for child molesters. Behaviour Research and Therapy, 26, 499-511. Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011 184 International Journal of Offender Therapy and Comparative Criminology Marshall, W. L., Bryce, P., Hudson, S. M., Ward, T., & Moth, B. (1996). The enhancement of intimacy and the reduction of loneliness among child molesters. Journal of Family Violence, 11, 219-235. Marshall, W. L., Eccles, A., & Barbaree, H. E. (1991). Treatment of exhibitionists: A focus on sexual deviance versus cognitive and relationship features. Behaviour Research and Therapy, 29, 129135. National Committee on the Provisions of Sex Offender Treatment. (1996). Standards and guidelines for the provision of services to sex offenders. Ottawa, Ontario: Correctional Service of Canada. Quinsey, V. L., Harris, G. T., Rice, M. E., & Cormier, C. A. (1998). Violent offenders: Appraising and managing risk. Washington, DC: American Psychological Association. Seto, M. C., & Barbaree, H. E. (1999). Psychopathy, treatment behavior and sex offender recidivism. Journal of Interpersonal Violence, 14, 1235-1248. Tabachnick, B. G., & Fiddell, L. S. (2001). Using multivariate statistics (4th ed.). Boston: Allyn & Bacon. Donna L. Mailloux, M.A. Department of Justice Canada, Research and Statistics Division 284 Wellington St. East Memorial Bidg., 6th Floor, Room 6264 Ottawa, Ontario Canada K1A OH8 Jeffrey Abracen, Ph.D. Central District Parole, Department of Psychology 330 Keele St. Toronto, Ontario Canada M6P 2K7 Ralph Serin, Ph.D. Research Branch, Correctional Service of Canada 340 Laurier Ave. W., Ottawa, Ontario Canada, K1A OP9 C. Cousineau Research Branch, Correctional Service of Canada 340 Laurier Ave. W., Ottawa, Ontario Canada, K1A OP9 Bruce Malcolm, Ph.D. Millhaven Institution P.O. Box 280 Bath, Ontario Canada K7L 4V7 Jan Looman, Ph.D., C. Psych. Department of Psychology, Regional Treatment Centre (Ontario) 555 King St. West Kingston, Ontario Canada, K7L 4V7 Downloaded from ijo.sagepub.com at QUEENS UNIV LIBRARIES on June 17, 2011