This executive summary describes results of a recidivism study of participants in the Sex Offender Treatment Program at Hiland Mountain Correctional Center from January 1987 to August 1995. The study was conducted by the Alaska Department of Corrections, Offender Programs, and the Alaska Justice Statistical Analysis Unit, Justice Center, University of Alaska Anchorage. Copies of the complete report Sex Offender Treatment Program: Initial Recidivism Study by Anthony M. Mander, Martin E. Atrops, Allan R. Barnes, and Roseanne Munafo (July 31, 1996) are available from the Department of Corrections through Rose Munafo at (907) 269-7416. A copy is also available for review in the Justice Center Library. A shorter version of this summary was published as an article in the Alaska Justice Forum.
Sex Offender Treatment Program: Initial Recidivism Study Executive Summary
The Alaska Department of Corrections, in conjunction with the University of Alaska Anchorage Justice Center, recently completed a study of sex offenders in the treatment program at Hiland Mountain Correctional Center during the period of January 1987 to August 1995. The study included analysis of descriptive characteristics of the participants; treatment variables such as length of time in program, reason for discharge and treatment stage at discharge; and re-offense data. The treatment group was compared with three other groups, including a motivated control group, an unmotivated control group, and a group of non-sex offenders (generics). There were several significant findings from the study:
•A treatment effect was clearly demonstrated. Treated sex offenders lasted longer in the community before they re-offended than offenders in any of the comparison groups. Even under varied definitions of re-offense, the treatment group lasted longer without re-offense regardless of the definition applied. (See Figure 1.)
•Treatment at any level improved survival in the community without re-offense. Those who were in treatment longer tended to last longer in the community without a re-offense. Those who completed all stages of treatment through the advanced stage had a zero re-offense rate for sexual re-offenses. This included Sexual Assault offenders (rapists), who generally tend to re-offend more quickly and at a higher frequency. (See Figure 2.)
•Sexual Assault offenders (rapists) do as well as Sexual Abuse of a Minor (SAM) offenders, both in terms of how long they stay in treatment and how far they advance through the program. This is an unusually positive outcome that has not been typically reported by other programs. For Sexual Abuse of a Minor (SAM) offenders, there was little difference shown in re-offense rate between those who reached intermediate phase and those who reached advanced phase in the program. For Sexual Assault (SA) offenders, achievement of intermediate or advanced stage was related to survival as measured by first arrest. (See Figures 3 and 4.)
•Alaska Native offenders do not progress as well in the program as non-Native offenders. This is the first study which directly addressed the demographic characteristics of Alaska Native sex offenders and the findings were somewhat different from what had been expected. It was originally anticipated that Alaska Natives who left the program early would be young, less educated and have a history of both alcohol and drug abuse. In fact, the study demonstrated that older, more educated Alaska Natives left the program early. The effect of a combined alcohol and drug abuse history was as expected. Two-thirds of offenders in the treatment group had a history of substance abuse, with an even higher incidence among Alaska Native offenders. Those who had no history of substance abuse tended to advance further in the program; those with a history of both alcohol and drug abuse tend to leave inthe earlier phases of the program.
Scope of the Study
This analysis of sex offenders who have received treatment through the Department of Corrections involved the creation of a workable database, complex data compilation and interpretation of statistical information. The study was limited in scope to the extent that only sex offenders who participated in treatment at Hiland Mountain SOTP were included. The study did not address sex offenders who received treatment through other Department sex offender treatment programs, either in an institution or community setting. It also did not address sex offenders who received treatment for other presenting problems, such as alcoholism or mental illness, exclusive of sex offender treatment. There are other limitations on the scope of the study addressed in the body of this summary.
DOC Sex Offender Treatment Program Description.
The sex offender treatment program (SOTP) has been developed over a number of years by the Alaska Department of Corrections (DOC), in conjunction with a variety of individual contractors. DOC has attempted to develop the programs along a continuum of care in a number of regions throughout the state. For a period of time in the evolution of sex offender treatment programs in Alaska, three distinct programs operated simultaneously. While there was some advantage to having treatment available in the three main regions of the state, consistency between the programs was an ongoing problem. One recommendation stemming from program evaluations conducted in FY 1992 was to create a continuum of services rather than duplicating services in several areas. It was also recommended that DOC adopt the Relapse Prevention model as its primary approach. This model has been shown to be particularly effective in the treatment of sex offenders.
Relapse Prevention Model.
Relapse Prevention (RP) is defined as a maintenance-oriented self-control program that teaches sex offenders how to determine if they are entering into high risk to re-offend situations, self-destructive behaviors, deviant cycle patterns, and a potential re-offense. This model, adapted by Pithers, Marques, Gibat and Marlatt (1983) from a substance abuse model developed by Marlatt and Gordon (1980), is a cognitive-behavioral approach to treatment. RP is based on the reality that although sex offenders cannot be "cured," they are responsible for their behaviors and can control them. It helps them explore factors which lead to committing sexual offenses and teaches them that they must make a commitment to abstain from participating in future deviant sexual behavior. The focus is not to "cure" or remove all temptation, but to develop ways to manage and cope with ongoing sexual desires, to teach the individual to be responsible to internal and external stressors (Salter, 1988). Prevention of sexually deviant, criminal, and other abusive and destructive behavior is promoted as the primary goal for all sex offenders who enter treatment.
Hiland Mountain Sex Offender Treatment Program.
The HMCC program is currently the only multi-phase institutional treatment program for sex offenders in Alaska. This program currently houses approximately 85 sex offenders in a milieu setting. Seventy of these are involved in intensive treatment programming and 15 are involved in pre-treatment programming/screening. Services are provided by a unique blend of contract therapists and specially-trained correctional officers. The SOTP consists of the following four program stages:
•Pretreatment: The purpose of this stage is to provide assessment, orientation, education, challenge of offense denial, and clinical management.
•Beginning Treatment: This stage prepares offenders to give and receive feedback, to use self-regulation and social skills, to assume responsibility for the current offense and its impacts upon victims, and, focuses on the most immediate precursors to the sexual offense with the creation of external management strategies. •
Intermediate Treatment: This stage addresses the earliest precursors to the offense and develops the skills for more self management of all risk factors. In the Intermediate phase the focus is on the internalization of skills learned in the preceding phase.
•Advanced Treatment: This stage emphasizes the application and generalization of skills to new situations.
With the exclusion of pretreatment, each stage is a minimum of six months and may take 12 months or more. Duration in treatment depends upon the offender's individual resources, problem areas, skills, motivation and length of sentence. The sex offender population is diverse, therefore, there are different levels of outcome. The HMCC SOTP is not designed with the expectation that every sex offender will complete all stages of treatment. Some offenders may leave the program without completing all stages. These offenders may lack the ability or the sentence length to go further in the program, but will have still gained some benefit from treatment when they leave the program. Regardless of which stage is reached, offenders are eligible for follow-up in community programs.
Groups Studied. The data from treated offenders in the present study was compared with data from sex offenders and non-sex offenders in several other comparison groups. This allowed some conclusions to be drawn regarding whether the results are likely to be due to treatment efforts or some other random and unknown factors. While the comparison groups are not ideal, their inclusion allows for more confidence in the conclusions. Data was collected on 685 convicted offenders, divided into the following groups:
•Sex Offender Treatment Group: 411 male sex offenders who received some level of treatment in the Hiland Mountain SOTP between January 1987 and August 1995. The amount of treatment varied considerably between individuals because of differences in length of sentence and speed with which progress was made by individuals in the program, however, all had some exposure to treatment.
•Motivated Sex Offender Control Group: 74 male sex offenders who requested treatment but did not receive treatment due to an insufficient sentence or the lack of an available treatment bed. This group of men was considered a "motivated" control group as they were willing to accept treatment but, through no fault of their own, this was not available.
•Unmotivated Sex Offender Control Group:100 male sex offenders randomly selected from the Offender Based State Corrections Information System (OBSCIS) who did not seek or request treatment. These are considered to be an "unmotivated" control group as they refused treatment, at least while in prison. Closer examination showed that some of the offenders randomly selected were already members of the treatment group, so that only 86 offenders were maintained in this group once the data was corrected.
•Non-Sex Offender Control Group: 100 male non-sex offenders who were randomly selected from OBSCIS to provide some comparison between sex offenders and the broader criminal population.
It is important to note that although these groups were not intentionally matched on demographic variables, a study conducted by Howard (1995) compared a sample of 358 of the offenders on several demographic variables. Results indicated no significant differences between the groups on any of these variables. We can, therefore, have some confidence that subjects in these two groups were at least roughly equivalent. The current research did not study all sex offenders who ever received treatment services from DOC. Not all sex offenders ever treated at Hiland Mountain have been included in the study, due to the fact that there is little or no data available on offenders in the program prior to 1985. Over the years, there have been a number of sex offenders who were treated at sites other than HMCC, including institutional treatment and pre-treatment at Lemon Creek in Juneau and Fairbanks Correctional Center. Additionally, a large number of offenders have received treatment in the various community settings throughout the state.
The amount of information available for analysis varied according to which of the treatment and control groups the offenders were in, as well as by the time periods when the offenders were in program. Sex offenders who were in the treatment group had a greater amount of information in their data file. Within the treatment group, more information was known about more recent program participants. The least amount of data was available for offenders in the control groups as there was no treatment file on these individuals. Information about re-offenses was derived from the Alaska Public Safety Information Network (APSIN).
Definitions of Recidivism.
Several measures of recidivism were used in the present study, including:
•First Arrest - Any Offense: This variable is a measure of both sexual and non-sexual re-offenses. The time it took for an offender to be arrested for any offense is reflected in this figure.
•Most Serious Offense - Any Offense: This variable is also a measure of both sexual and non-sexual re-offenses but specifically determines the most serious of all re-offenses committed by an offender. This was determined by looking at NCIC offense codes and applying an algorithm to identify seriousness. The algorithm used was developed by the Federal Bureau of Justice Statistics in the mid-1970's in an attempt to arrange the NCIC codes according to level of seriousness.
•First Arrest - Non-Sexual Offenses: This variable is a measure of re-offense for any non-sexual crime.
•Most Serious - Non-Sexual Offenses: This variable is a measure of the most serious of the non-sexual re-offenses, which is assessed using the algorithm described above.
•First Arrest - Sexual Offenses: This variable separates sexual offenses from other offenses so that we can study the effects of treatment on sexual re-offending specifically.
•Most Serious Sexual Offense: This variable examines the most serious of the sexual re-offenses using the same algorithm as described above.
There is a range of criminal behavior reflected in the above definitions. Measures which reflect criminal behavior of any type tend to be the most sensitive since they pick up criminal thinking of any kind. Sexual re-offenses are the least sensitive since they are typically under reported. Non-sexual offenses, however, are related to sexual offenses because sexual offenses are often at the end of a chain of events which include non-sexual precursors. It is this chain of events which the relapse prevention plan addresses. Statistical Analysis. The statistical procedure used to analyze re-offense data in the present study is called "survival analysis". This procedure, discussed by Marques, et al. (1994), accounts for the differential time that offenders are in the community. The longer offenders are in community placement the greater the opportunity for re-offense The method, therefore, takes into consideration the fact that offenders have varying opportunity to re-offend. Survival analysis is commonly used in medical research. The procedure yields a survival curve whereby groups can be compared for survival over time. It has the additional benefit of accounting for all offenders who have been released, regardless of the length of time that has elapsed since release. The most effective treatment methods will result in a longer period of surival without re-offense. In the present study, we found that treated offenders survived at a higher rate than those offenders who were not treated. There was also a great deal of demographic information, as well as information relating to the treatment process, which was analyzed. Whenever possible, this data was analyzed using appropriate statistical methods (e.g. chi-square) to determine if differences were statistically significant. In some cases, statistical analysis was not possible due to insufficient numbers of subjects.
This research demonstrates that treatment can and does work, certainly for some offenders. It works by reducing the incidence of sexual re-offense or by prolonging the time until re-offense. Either of these results reduces the number of victims in the community. When treatment does not work for certain offenders this information is equally important. Sexual assault is not a disease that can be cured. It is an aggressive deviant behavior that results from the convergence of a complex number of factors. It can, however, be contained and managed. Offenders who are amenable to treatment and willing to actively participate learn to recognize precursors to relapse and self-manage their high risk behavior. Those who are not amenable and/or not willing to participate in treatment must be controlled by external measures. It is important to recognize that offenders differ along a continuum of risk. Identifying the extent of the risk and the conditions under which an offender is likely to relapse allows the offender and others to manage the risk more effectively. It is possible, through continued research effort, to determine a constellation of factors which would predict treatment success at different levels. DOC should attempt to identify such predictors through continued research. This would assist in developing a more objective definition of "maximum treatment benefit." We cannot expect all offenders to reach the advanced treatment stage, but it would be helpful to have a more objective criteria for determining when offenders had derived, what is for them, the maximum benefit. External supervision could then be altered according to the level of risk. This would partly be determined by treatment stage and partly by other variables such as seriousness of offense and other factors. Such an approach would increase the efficiency of treatment. It would also provide a more objective basis for making decisions about furlough and parole, thereby enhancing community safety.
Many of the findings suggest that Alaska Native offenders leave treatment early. The reasons for this are not entirely clear. Overall, Alaska Native offenders in the Hiland Mountain Sex Offender Treatment Program tend to be younger, abuse substances more than non-natives, and have less formal education. In the overall population of the program, each of these factors have been shown to contribute to the lack of advancement in treatment in the present study. However, other analysis suggested that it was older, more highly educated Alaska Natives who left the program early. Severe substance abuse also appeared to be related to early discharge. This is the first study of an Alaska Native sex offender population and it is limited in scope to treatment program participants. There have been very few comprehensive studies done on the characteristics of sex offenders in indigenous populations in the United States, although some research has occurred with these populations in other countries, such as Australia and New Zealand. This research appears comparable to these studies, both in terms of the size of sample studied and the length of time studied. Because a baseline has been established for this group, it will enable evaluation of progress as the program develops. Further research is needed to identify and correct the specific problem(s) leading to early withdrawal from program.
During the past two years, the HMCC program has undertaken a number of steps to improve services to Native offenders. Changes have included a pre-treatment orientation for Natives, mentoring by more advanced-phase Alaska Natives, monthly Alaska Native Planning and Support Group, staff training and curriculum development to incorporate Alaska Native values and practices, collateral contacts with families, initiation of a "week of reflection" which encourages offenders to reflect about their desire to quit program and other changes to try to make programming more relevant and appropriate to Native offenders. Additionally, efforts are currently underway to further augment pre-treatment and beginning treatment by incorporating Native healing and spiritual practices. The data included in this study only includes participants through August 1995. Given these changes, it will need to be determined if the same pattern emerges, or if there is a change, once the current year's data is analyzed.
Conclusions and Recommendations
There are many reasons for funding effective sex offender treatment that reduces crime and the number of victims, including some practical reasons. Several studies have demonstrated how effective sex offender treatment saves taxpayer dollars (Prentky and Burgess, 1990; McGrath, 1994; Steele, 1995). If one assumes that there will frequently be more than one victim, both pre- and post-imprisonment, then it follows that the cost of incarcerating sex offenders without treatment will be higher than that of treating them. Following this logic, these studies appear to demonstrate that treating sex offenders is more cost effective than incarceration without treatment. The nature of sex offenders necessitates close teamwork and cooperation. Sex offenders are adept at finding the weak link in the system and using it to their advantage. All parts of the system must work as a team for treatment to be successful. Based on the results of this research, it is recommended that the Department of Corrections develop a plan for addressing the system-wide issues involved in the treatment of sex offenders in its custody, both within the correctional institutions and on probation/parole supervision. At a minimum, the following should be addressed:
•Determine changes in service delivery which may be needed in order to address the differences between Sexual Assault (SA) and Sex Abuse of a Minor (SAM) offenders.
•Assess the needs of Alaska Native sex offenders and determine what effect changes implemented within the past year have had on this population.
•Develop a plan for future research on the factors that can predict treatment success.
•Identify system gaps and develop a plan for future development of both institutional and community programs for sex offender treatment and management.
•Develop standards for external monitoring of sex offenders who are determined to be unamenable to treatment.
•Develop recommendations for changes needed in other branches of government, including possible statutory changes. Identify potential funding for further research and program development.