James M. Byrne, Ph.D., The University of Massachusetts, Lowell
Faye S. Taxman, Ph.D., Virginia Commonwealth University
ACCORDING TO A recent Bureau of Justice Statistics (BJS) review of reentry trends in the United States, there were 1,440,655 prisoners under the jurisdiction of federal or state correctional authorities at year-end 2002 (Hughes and Wilson, 2003). During the year, there was a constant flow of offenders both into prison (close to 600,000 individuals) and out of prison (again, about 600,000). Offenders entering prison were either newly sentenced offenders (60 percent) or parole/other conditional release violators (40 percent). Offenders leaving state prison included drug offenders (33 percent), violent offenders (25 percent), property offenders (31 percent), and public order offenders (10 percent). About one in five of these reentry offenders were released unconditionally; the remaining offenders were placed under parole supervision. Overall, it is projected that 67 percent of these releasees will likely be rearrested and 40 percent will likely be returned to prison within three years of their release date, based on a recent BJS study (Langon and Levin, 2002). Clearly, a subgroup of the federal and state prison population appears to have integrated periods of incarceration into their lifestyle and life choices. The constant movement of these offenders into and out of prison has negative consequences not only for offenders but also for the community at large, including victims, family members, and community residents. What can and should the corrections systems do to "target" these offenders for specialized services and controls to improve reintegration into the community?
In the following article, we examine the offender targeting issue in detail, utilizing data gathered from our review of eight model Reentry Partnership Initiative Programs1 (see Taxman, Young, Byrne, Holsinger & Anspach, 2003 for an overview of research methodology). We begin by describing the changing patterns of federal and state prison admissions and releases. We then examine the target population criteria used in the eight model RPI programs and discuss the unique challenges presented by different offender groups, including repeat offenders, mentally ill offenders, sex offenders, and drug offenders. We conclude by identifying the relevant classification, treatment, and control issues that decision makers will have to address as they design and implement their own reentry processes for targeted offenders and/or communities.
1. Reentry Trends: Changing Patterns of Prison Admission and Release
The number of prisoners under state and federal jurisdictions has increased dramatically over the past eight decades. In 1925, there were 91,669 state and federal prisoners and the rate of incarceration was only 79 per 100,000 of the resident population. By the end of 2000, the number of incarcerated offenders rose to 1,321,137, which translates into a rate of incarceration of 478 per 100,000 residents. The change in the correctional landscape followed the shift in sentencing philosophy from rehabilitation to incapacitation, which grew out of frustration with offenders who refuse to change, the failure of rehabilitative programs to reduce recidivism, and the need to punish offenders for their misdeeds. Paradoxically, the incapacitation approach has resulted in more institutionbased punishment for offenders, but less community-based control of the returning home population.
Offenders are released from prison either conditionally or unconditionally. For the three out of four offenders released from prison conditionally in 1999, a supervised, mandatory release mechanism was used for 50.6 percent, some form of discretionary release via parole was used for 36.1 percent, and probation/other supervision was used for 13.3 percent. The remaining prison releasees —representing almost a quarter of the total release population (109,896—22.2 percent of all releasees) were sent back to the community "unconditionally," with no involvement of the state or federal government in overseeing their return to the community. That is, some type of supervised release (e.g., probation, parole, etc.) was not part of the reentry process. In the vast majority of these unconditional release cases (95 percent), the offender was released from prison due to an expiration of sentence.
Any discussion of the impact of our returning prison population on community safety must begin by recognizing the fundamental changes in release policy in this country over the past decade. Supervised mandatory release is now the most commonly used release mechanism by state prison systems, while the vast majority of federal offenders are released upon expiration of their sentence. Focusing for a moment on regional variations in release policy, we find that prison systems in the Midwest (35.4 percent of all releasees) and Western states (77.2 percent of all releasees) are more likely to rely on the supervised mandatory release mechanism than on either expiration of sentence or discretionary parole release. In the Northeast, the pattern is noticeably different: discretionary parole release is the most common release mechanism in these states (60 percent of all releases). This was also the pattern found in Southern states, although there is clearly a lower rate of discretionary parole releasees (33 percent of all releasees) and more use of expiration of sentences (30 percent of all releasees) and/or supervised mandatory releasees (22 percent of all releasees) in this region.
Despite the growing trend toward the use of mandatory release mechanisms and away from discretionary parole release, we should emphasize that several states (21 in 1997) do not use this release mechanism at all. Six of them (Maine, Massachusetts, Ohio, Delaware, Florida, and Nevada) relied more often on expiration of sentence than on any other release option and in four of these states, supervised mandatory release was not available. Due to changes in parole practices, parole boards are reluctant to release offenders early. The growing trend is for more offenders to be released with either minimal time under community supervision, or without any community supervision condition at all. While some scholars observe that many offenders are better off without community supervision, due to the problem of technical violations and recycling of offenders from prison to community to prison (Austin, 2001), others observe that more supervision is required to manage the reintegration process and to reduce the potential harm that offenders released from prison and jail present to the community (Petersilia, 2000; Taxman, et al., 2002). More research is needed in this area to determine the degree and level of supervised release (if any) that is useful to maximize community safety, but it certainly appears that changes in sentencing policies and release practices have likely had negative consequences for offenders and the communities to which they return.
Since many states have opted not to develop policies and procedures to allow supervised mandatory release, it is likely that more and more offenders will be "maxing-out" of prison in these jurisdictions. Do these offenders pose a greater threat to community safety than either the parole or mandatory release population? A recent study by the Bureau of Justice Statistics found that mandatory parolees are less likely to successfully complete parole than discretionary parolee discharges (Hughes, Wilson, & Beck, 2001). While we do not know the answer to the question about the relative effectiveness of different release mechanisms, it is important to continue to monitor this issue.
We do know that offenders are now serving a greater proportion of their sentences in prison and regardless of the method of release, they are returning to the community with the same problems (e.g., lack of skills to obtain employment, substance abuse problems, family problems, individual mental health and physical health problems, repeat offending behavior, etc.) that they had when they were first incarcerated (Maruna and Immarigeon, 2004). In addition, some offenders are returning to the community with new mental health, physical health, and personal (criminogenic) problems, due to such factors as negative institutional culture (Bottoms, 1998; Sparks, Bottoms and Hay, 1996), increased incarceration period (Austin, 2001), the spread of communicable diseases in prison (Rand, 2003), and isolation from the community (Maruna, 2004). While they were incarcerated, the communities they used to reside in may have improved (due to such factors as community mobilization and betterment activities, a better economy, community policing, etc.) or they may have deteriorated (due to economic downturns, increased gang activities, the spread of infectious disease, etc.). In either case, the community prisoner's return may be to quite a different community from the one they left. The longer offenders remain in prison, the more likely that there will be changes in family, peer associations, and neighborhood dynamics needing to be addressed during reintegration. All of these changes complicate reintegration, but they must be considered when designing and implementing offender reentry programs. As Gottfredson and Taylor (1986) demonstrated almost two decades ago, these person-environment interactions likely hold the key to understanding (and changing) the behavior of offenders released from prison.
2. Offender Targeting for Reentry: An Overview of Current Practices
The Office of Justice Programs (OJP), in conjunction with a wide range of federal agencies involved in offender reentry directly or indirectly, has recently allocated 100 million dollars to help fund reentry initiatives in every state and U.S. territory, including Puerto Rico and Virgin Islands. Beginning in 2002, 68 separate reentry programs have been designed, developed, and implemented, targeting a diverse group of juvenile and adult offenders. However, a recent BJS review of reentry trends revealed that in 2001, nearly half of all state prison releasees were from the following five states: New York, California, Illinois, Texas, and Florida. Table 1 provides an overview of the OJP programs in these five states, focusing on program size, location, and initial targeting criteria. It appears from our preliminary review of these programs that the OJP initiative will likely include only a fraction of these states' releasees, which makes the decision on whom to include and whom to exclude even more critical. Unfortunately, a detailed review of the initial development of the OJP reentry initiative has not been completed, although the Urban Institute has been selected to conduct the initial evaluation of this program.2 In the interim, we are left to sort through a large number of program descriptions (see OJP's web site for state-specific descriptions of reentry initiatives) and a small number of case studies and process evaluations.3 Despite this evaluation research shortfall, it certainly appears that governors, legislators, and corrections administrators are jumping quickly onto the reentry bandwagon. The question we focus on in this article is straightforward: who (and where) should we target for reentry? To answer this question, we have examined the targeting criteria developed in eight model reentry partnership initiatives (RPI) programs identified by the office of Justice Programs and recently included in a detailed multisite process evaluation conducted by the University of Maryland's Bureau of Governmental Research (for an overview, see Taxman, et al., 2003). It is our view that the targeting issues identified in the following review of the eight RPI programs4 will be applicable to 68 new reentry initiatives currently in development across the United States. In general, the reentry programs we reviewed can be described as including three separate reentry phases: 1) the institutional phase, 2) the structured reentry phase, and 3) the community reintegration phase. However, considerable variation not only in the design, but also in the duration of each of these reentry phases appears to be related—in large part—to the specific targeting decisions made by program developers at each site. In the following section, we highlight the impact of offense, offender, and area-specific targeting decisions on each phase of reentry.
A. Targeting and the Institutional Phase of Reentry
Our review of RPI programs found considerable variation in what actually constitutes the "institutional" phase of offender reentry. In one jurisdiction (Burlington, Vermont), offenders were identified and selected to participate in the reentry program upon entrance to prison, during the initial prison classification process. In the remaining jurisdictions, identification of potential reentry participants occurred several months prior to the inmates expected release date (6 months to 1 year). Obviously, this basic decision has important ramifications for both the offender and the institution, particularly when participation in specific prison-based treatment programs is a feature of the reentry program. Regardless of when this phase of reentry began, it appears that inmates participating in the RPI programs we reviewed had access to programs and services not available to other inmates at these facilities. In this respect, treatment availability, access, and perhaps even quality represent important advantages linked to participation in the reentry programs we reviewed.
In the institutional phase of the reentry process, offenders who meet the RPI site's target population criteria are initially identified and contacted about possible participation in the reentry program. For offenders being released unconditionally, program participation is voluntary; however, conditional releasees may be required to participate as a condition of parole. Program developers at prospective RPI sites are faced with several difficult decisions regarding initial offender targeting. First, due to program size restrictions, RPI model programs at the sites we reviewed targeted specific release locations for reentry. Second, only a subgroup of all offenders to be released to these locations is usually targeted for potential reentry participation. Third, targeting may vary not only by location and offense type but also by the method of release (i.e., conditional vs. unconditional). And finally, program participation may be restricted to offenders who are at a certain level of institutional control (e.g., medium security), due to size limitations and/or institutional control concerns.
Regardless of exactly how the final group of RPI program participants is selected, the institutional phase is expected to include a range of offender programming options designed to prepare offenders for resuming their lives in the community. These program options would likely include education, vocational training, life skills, and of course, individual/group counseling. In three sites, the emphasis was on providing motivational readiness for treatment, in order to prepare the offender to make significant lifestyle changes as they return to the community. As we have noted in a separate report (see Taxman, et al., 2003), we maintain that reentry programs should be oriented toward preparing inmates for return to the community from the outset of their institutional stay. However, only one of the eight RPI models we visited (Burlington, VT) began the institutional phase during the first several months of an offender's incarceration. A much more common approach is to begin the institutional phase of the reentry program several months before the offender's targeted release date, but prior to the pre-release phase. In fact, several of the RPI programs we reviewed had the institutional phase folded into the structured reentry phase, making it difficult to determine where one phase ended and the next began.
B. Targeting and Structured Reentry
Structured reentry is the catchphrase for perhaps the most critical step in the offender's reentry process. During structured reentry, the offender must make the transition from institutional to community control. In the programs we reviewed, structured reentry began approximately 1 to 3 months prior to the offender's targeted release date and continued through the end of the offender's first month back in the community. It consisted of two distinct but interrelated stages (the in-prison and in-community stages), which were designed to offer a seamless system of transition from prison to community.
The structured reentry process requires coordination and collaboration between and among several distinct "partners" in the reentry process, including the offender, victim, community, treatment providers, police, and institutional and community corrections. As we have already observed regarding the institutional phase, "structured reentry" will likely be a different experience for offenders released conditionally than for those offenders (about 20 percent of all releasees nationally) released unconditionally. However, the components of structured reentry likely will require the development of a plan for each returning offender targeted for participation, focusing on such basic issues as: 1) continuity of treatment, as offenders move from institutional to community treatment providers and address longstanding criminogenic factors (e.g., substance abuse, mental illness, repeat offending, etc.); 2) housing options; 3) employment opportunities; 4) family needs and services; and 5) victim/community concerns (e.g., safety, restitution, public health, reparation).
Some jurisdictions (i.e., Florida, Maryland, and Nevada) found it advantageous to move offenders closer to their release location during their last few months in prison to facilitate the community reintegration process. In theory, locating the offender closer to home should help him or her to renew family ties, obtain employment and secure appropriate housing and treatment. We suspect that these kinds of community linkages may actually be more important for offenders released unconditionally, without the specific forms of community treatment, supervision and control associated with the typical offender conditional release plan. For both conditional and unconditional releasees participating in a reentry program, it appears that some form of offender movement may be needed during the structured reentry phase, particularly if participation in a specific treatment program is a component of the reentry program and linkages need to be established to ensure provision/ continuity of treatment.
C. Targeting and Community Reintegration
Phase III of the reentry programs we reviewed is referred to as the community reintegration phase. For many offenders leaving prison, the initial period of adjustment (i.e., the first one or two weeks after release) is actually less difficult than the subsequent period of community reintegration (see, e.g. Taxman, Young, and Byrne, 2003). There are a variety of possible explanations for this phenomenon. First, keep in mind that essentially two groups of offenders are being released from prison: conditional and unconditional releasees. While both groups of offenders are offered similar support services (e.g., employment assistance, housing assistance, health care and treatment), conditional releasees are monitored by community supervision agents who have the power to revoke their parole if they refuse this "assistance." With the notable exception of sex offender registration, no such controls can be invoked for the unconditional releasee population, although the RPI initiative has pioneered the use of a number of informal social controls to induce offenders to fully participate in the reentry program. These informal social controls include the use of guardians and advocates in the community, who are available to assist the offender with reintegration, helping the offender make linkages with services, employers, and community groups (such as faith-based, self-help groups, etc.). The relationship that develops between guardians and returning offenders may have a positive influence on program participation and compliance.
It is certainly possible that after an initial period of compliance and participation, offenders from both groups will begin to return to earlier behavior patterns, such as gang participation or drug/alcohol abuse. For offenders under conditional release status, the use of behavioral contracts with clearly defined rewards and sanctions may reduce the number of offenders who backslide in this way. However, successful reentry programs must develop alternative mechanisms for fostering compliance among offenders released from prison unconditionally. For example, one site we visited proposed making "housing" assistance available to offenders actively participating in the reentry program. Stated simply, an offender may be released unconditionally from prison, but his or her participation in the reentry program is conditional on compliance with the program's rules and regulations (such as no drugs or alcohol, curfews, participation in treatment, etc.). If an offender wants to live in housing provided by the RPI, then he or she will continue to participate in treatment, remain employed, etc. In one RPI model we reviewed, housing is provided for up to 90 days. However, the program allows the offender to live in transitional housing for an additional 90 to 370 days for a minimal fee as the offender becomes stabilized in the community For many offenders, housing may represent a more effective inducement than the threat of other sanctions (Taxman, Young, and Byrne, 2003).
3. Variations in Targeting Criteria for Reentry
Any discussion of offender reentry must begin by recognizing that urban areas, often with high concentrations of minorities, are "home" to the vast majority of returning inmates in the United States. Approximately 600,000 prison inmates returned to the community in 2002 alone (Hughes and Wilson, 2003); over half of these returning offenders were from five states (California, Florida, Illinois, New York, and Texas). To many observers, the answer to the question "whom should we target for reentry?" is straightforward: all releasees from our state and federal prison system, regardless of location, release status, conviction offense type, and/or criminal history. However, an examination of the target population criteria used to select offenders in the eight model RPI programs we reviewed presents a more pragmatic, stakes-oriented view of the targeting issue: do not place "high stakes" offenders (such as sex offenders) into a new reentry program.
This approach clearly fits the cardinal rule of correctional practice: inaugurate new initiatives by focusing on offenders who are likely to be compliant and less likely to create public outcry. The "low-risk/low-stakes" approach is promoted as a means to build community and stakeholder support for new concepts with the expectation that, if the innovation is successful, then corrections officials will expand the target population. In fact, many criminologists continue to argue that we are likely to see the largest reductions in offender recidivism when we target the highest-risk groups of offenders for program participation (Taxman, 2002). However, program developers may be less interested in recidivism reduction and more interested in the level of re-offending by program participants. When viewed in this light, the question becomes: how much recidivism is one willing (or able) to tolerate among offenders targeted for reentry?
Table 2 presents the results of our multisite review of target population criteria. Four of the eight programs we examined place offense restrictions on offenders considered for participation in the jurisdiction's new reentry program. All programs with offense restrictions specifically excluded sex offenders, utilizing information from both the offender's incarceration (or instant) offense and the offender's criminal history to identify ineligible offenders. In addition to restrictions on sex offenders, one jurisdiction places restriction on violent offenders, while another does not allow offenders who have ever committed a crime against children to participate. Another criterion used by staff at two sites was the psychological health of the offender. Offenders with a history of mental illness/ psychological disorders are excluded from participation at these reentry sites. According to a recent review by Liebling (1999 as cited in Petersilia, 2000), approximately 1 out of 5 prison inmates report having a mental illness (see also Lurigio, et al., this volume). Given the overlap among violent offenders, sex offenders, and mentally ill offenders, it appears that some RPI program developers used a multiple, offender/offense-based scheme to cast as wide a net as possible over the pool of multiple-problem offenders to exclude from the reentry programming.
We should emphasize that these ineligible offenders will still return to the community upon release, but they will do so without the specific support and control offered through the RPI effort at the eight sites we reviewed. Since a significant number of the unconditional release population who are "maxing out" of prison are sex or violent offenders, it appears that the very group of offenders raising the most community concern tends to receive the lowest level of correctional supervision and support. The paradox inherent in this decision is that it is precisely the group of offenders being excluded from reentry programs that would most likely benefit from participation in the programs, and that may present some of the greater public safety risks. Recent evaluation findings continue to demonstrate that larger gains in reducing recidivism are likely to occur with high-risk offenders who have a greater likelihood of committing new offenses (Andrews & Bonta, 1996; Taxman, 1998). As the RPI program grows and evolves, it is likely that many of the sites will expand the offender pool to include "high stakes" offenders. Three of the eight sites we visited understood this issue well enough to place no offense restrictions on reentry offenders for their specialized initiatives. In these jurisdictions, the key criterion was location. Reentry program developers reserve the reentry initiative to offenders returning to specific neighborhoods, regardless of their prior offense history, seriousness of current offense, or special needs (e.g., substance abusers, mentally ill).
4. Offender-specific Reentry Strategies: What Works, with Whom, and Why
As part of developing the RPI initiative, each jurisdiction had to consider the state of knowledge about the reintegration "challenges" posed by a wide range of institutionalized offenders. Decisions made about whom (and where) to "target" for specialized reentry programming will affect the structure and purpose of the RPI model being developed. As we have reported here earlier and in separate reviews (see Taxman, et al., 2003) high risk offenders, particularly sex offenders (however the pool is defined) have been excluded from participating in five of the eight model reentry programs we reviewed. Of course, sex offenders and other excluded offenders in these jurisdictions are still returning to the community, either on conditional or unconditional release status; they simply do not have access to the model programs, staff, services, and support that are being designed to maximize public safety. While such offensebased exclusion criteria may make sense to policy makers and program developers, they may actually make reintegration more difficult for "excluded" offenders.
What do we currently know about different types of offenders who will be returning to the community from our state and federal prison systems? The national statistics do not illustrate the tremendous variation in characteristics of offenders that occur by state and region. For program planners and developers, it is critical to examine state-specific (as well as region-specific) information about the characteristics of institutionalized offenders, and to design and implement reentry programs that are appropriate to the particular target population (offense type, offender type, demographic profile) and target area (e.g. urban, rural) included in the reentry initiative. In the following section we describe the unique reentry challenges presented by four groups of offenders: sex offenders, repeat offenders, substance abusing offenders, and mentally ill offenders. Although our focus is on offenders, we recognize that communities also will vary (e.g. tolerance, support, resources) in ways that will affect the offender reentry process.
A. Sex Offenders. What is a sex offender? To many observers, the answer is obvious: anyone convicted of a sex-related crime. In 1997, for example, there were 1,046,705 offenders in our state prison system: 2.6 percent of these incarcerated offenders were convicted of rape, while another 6 percent were convicted of some other form of sexual assault. By comparison, only a fraction of the federal prison population (8 percent of 88,018 federal prisoners) would be classified as sex offenders. Many offenders currently in prison for other crimes have a criminal history that includes at least one sex offense conviction as an adult, and an unknown number of our state prison population have a juvenile record for sex offending. Taken together, approximately one in five offenders returning from state prison facilities to the community each year could be categorized as sex offenders. The sex offender category consists of a variety of behaviors that include, but are not limited to rapists, child molesters or pedophiles, exposures and other sexual deviancy. These groups of offenders pose a major classification, treatment, and control dilemma for public and community safety officials attempting to address the offender reentry issue. The different types of behavior imply different levels of treatment and control that are needed to address public safety issues (CSOM, 2001).
It needs to be pointed out that, as a group, sex offenders have comparatively lower recidivism rates than either drug or property offenders. However, untreated sex offenders have re-offense rates that are twice as high as sex offenders who receive some form of treatment (see e.g., Alexander, 1999). Given these research findings, it is critical that RPI program developers think creatively about how to increase the treatment participation rates for sex offenders during all three of the reentry program's treatment phases. In those programs that exclude sex offenders, efforts will need to be made to explain the rationale for exclusion to residents of targeted communities. It is to be hoped that program developers in these jurisdictions will have more to fall back on than transfer mechanisms (via sexually dangerous offenders statutes) and sex offender registration.
B. Repeat Offenders. Repeat offenders are those offenders who have a history of criminal behavior, including offenses that affect the quality of life of the communities. Nearly 60 percent of the federal prisoners and 83 percent of the state prisoners have at least one prior criminal conviction (Beck & Harrison, 2001). The classification issues for repeat offenders present enormous challenges to reentry planners. The repeat offender has violated the norms of the community, whether for a serious (e.g., murder, rape, robbery, assault, etc.) or minor (e.g., public disorder, etc.) offense. When viewed in the broader context of criminal "careers" (or crimes in the life course), an offender's current offense tends to be rather misleading, because it does not detail the offender's criminal history or the pattern of criminal behavior. [see table 3] For example, more offenders are in prison for public disorder offenses (up 114 percent in ten years), but their incarceration is more likely due to their criminal history than to the nature of the instant offense. Lynch and Sabol (2001) note that offenders in prison for violent offenses often have mental health and personal/behavioral issues that are not addressed in traditional prison treatment programs. For this reason, they are more likely to have behaviors that will carry over into the community. As a recent nationwide review of prison classification systems demonstrated (Hardyman et al., 2004), few prisons have a classification system that adequately assesses the offender's criminogenic risk and need factors; therefore leaving reentry planners without a good understanding of the psychosocial functioning of the offender at the time of release from prison. Further, the complexity of criminal "careers" typologies generally (and offender crime trajectories in particular) underscores the need to identify subgroups of offenders at different stages in their criminal careers.
A related category is the churners, or offenders that are in the prison-parole-prison cycle due to technical violations or new arrests while on supervision. As reported by Lynch and Sabol (2001), 36 percent of the prison releases in 1998 were prisoners who were released from a subsequent prison term on an original sentence. In other words, they had been in prison and released and then returned to prison for "mishaps" in the community. These repeat offenders present public safety challenges because they have already been unsuccessful in their reintegration; for these offenders, reentry applies equally to prison and community, which suggests the need to rethink our approach to this group of releasees.
C. Substance Abusers/Drug Offenders. Most correctional administrators readily recognize that most offenders are substance abusers, with national surveys noting that 80 percent of the state prisoners and 70 percent of the federal prisoners self-reported past drug and/ or alcohol use (Mumbo, 1999) and 16 percent reported committing the current crime to obtain funds for illicit drugs. A 1997 study of prisoners used the CAGE, a standard protocol in the field of alcohol assessment, finding that 24 percent of the offenders are alcohol dependent. The study did not use techniques to estimate the drug dependent population. However, as part of the Arrestee Drug Abuse Monitoring (ADAM), researchers found that nearly 80 percent of the offenders report past drug and/or alcohol use and 51 to 79 percent of the arrestees (with a median of 65 percent) have positive urinalysis at the time of the arrest for marijuana, cocaine/crack, heroin and opiates, methamphetamines, phencyclidine (PCP), and benzodiazepines. Using the DSM IV criteria to define drug dependency, 38 percent of the offenders were found to be dependent and in need of treatment. Actually, approximately half of the positive offenders (34 percent of all offenders) were considered heavy drug users based on the commonly accepted criteria of using drugs for at least 13 (Taylor, et al., 2001). Findings from ADAM indicate the offenders are not homogenous in their use patterns and in fact there is tremendous variation in their use of illicit substances. Marijuana, in most regions, continues to be the drug of choice. While offenders tend to test positive for one drug, behaviorally the dependent user tends to use an array of illicit substances depending on the availability.
The challenge for correctional officials is to identify the subgroup of returning offenders whose drug-alcohol dependency is directly associated with other forms of criminal behavior. This is the group that would most likely benefit from treatment and who pose a more immediate threat to public safety/ security. Other drug users—entrepreneurs or recreational drug users—are less likely targets for substance abuse treatment programming, since their criminal behavior is not affected by their drug use (Chaiken & Johnson, 1988). A challenge exists to identify offenders who have substance-abusing behavior that increases their risk-taking in their criminal activities. Correctional administrators and treatment providers must develop a classification scheme that distinguishes between the substance abusers and the criminals. Such a scheme will allow RPI stakeholders to target offenders to appropriate treatment services, based on the need for cost-effective reentry strategies. As a number of researchers have suggested, different strategies must be developed based on an understanding of the specific stage of an offender's addiction career and his/her criminal career.
D. Mentally Ill Offenders/Dual Diagnoses. Between 15 and 20 percent of the state prisoners have mental health issues that affect their normal functioning. A recent study of prisoners found that 14 percent had a mental health or emotional crisis in prison or were required to be admitted overnight. Nearly 10 percent of the offenders were using psychotropic medications within prison for their mental health issues, although six states had 20 percent of the offenders using medications (Hawaii, Maine, Montana, Nebraska, Oregon, and Vermont). A recent BJS study found that mental heath services are commonly provided in maximum/high-security confinement facilities. Further, nearly 13 percent of the state inmates receive some type of mental health therapy, usually counseling (Beck & Maruschak, 2001). The prevalence of mental health disorders among the prison population carries over to the community, where medication and access to services are limited due to lack of health insurance. The needs of mentally ill offenders are just becoming more apparent as mental health issues are identified, especially related to the specific problem of providing treatment (both institutional and community-based) for the multiple problem offender (see, e.g. Lurigio, et al., this volume). We know, for example, that mentally ill substance abusers fail in traditional substance abuse treatment programs at a significantly higher rate than other substance abusers. It is likely that similar failure patterns can be identified for the treatment of other offender problems (e.g., mentally ill sex offenders, mentally ill repeat offenders) that suggests that the multiple problem offenders require a different approach. Excluding mentally ill offenders and/or multiple problem offenders from the latest wave of reentry programs is obviously not the answer. Only a collaborative response from both mental health and corrections professionals (again, see Lurigio, et al., this volume) can begin to address this target population's problems and needs.
5. Conclusion: Offenders, Communities and the Need for Change
Our review of the targeting criteria used in eight model reentry programs (in Massachusetts, Vermont, South Carolina, Florida, Washington, Nevada, Maryland, and Missouri) highlighted a number of critical issues that must be addressed by program developers. First, it appears that program developers —while agreeing that the reentry process includes three distinct, but interrelated phases (institutional, structured reentry, and reintegration)—are having difficulty clearly putting into operation each of these three program components. This problem is most noticeable in our review of the institutional phase of reentry; it is often unclear when this phase begins, how it should be distinct from the normal institutional programming and processes, and where the initial institutional phase of reentry should be located (e.g., should all reentry offenders reside in the same facility, on the same wing, etc., utilizing a therapeutic community model? Or should they simply have access to the same programs and resources, regardless of location?).
A second question that emerges from our review is: Who should be targeted for reentry? As we highlighted earlier, only a small proportion of all offenders currently in prison will be released to one of the specialized reentry programs described here. A closer look at table 1 underscores this point dramatically. According to a recent BJS review of reentry patterns, half of the 600,000 adult prisoners released from state prison in 2001 came from the five states included in this table (California, Florida, Illinois, New York, and Texas). However, the target populations identified for these five states included fewer than one thousand prisoners, which represents less than 1/3 of 1 percent of all releasees from these five states. These selected inmates will be placed in programs funded by OJP's (100 million dollar) reentry initiative; about 9 million dollars was allocated to these five states for adult reentry programs, approximately $9,000 per offender released. Since only a fraction of the reentering prison population can be placed in this new wave of reentry programs, program developers need to consider carefully the criteria for selection they will utilize in their own jurisdictions.
Given limited resources, it certainly makes sense to begin by targeting specific locations for participation, perhaps based on an examination of previous release locations for each state's prison population. However, the danger inherent in restricting access to those offenders returning to a particular community or neighborhood is that 1) the program may actually increase the concentration of ex-offenders in a particular neighborhood, and 2) the identification of returning offenders in a specific area as reentry participants may result in the use of profiling strategies by local police (see Byrne and Hummer, this volume, for a discussion), which could have negative consequences for both offenders and communities.
Our review also revealed that program developers—often with a limited number of program placements—may exclude both high risk (to reoffend) and high stakes (to the program's legitimacy) offenders from the initial target population. The problem inherent in this strategy is that it excludes from participation the very group of offenders most likely to need the services, support, and control provided by the reentry initiative. Since these offenders will be returning to the community anyway, program developers may want to consider the possibility that by expanding their target population, they may actually improve both community satisfaction and community safety (if the program actually delivers on its recidivism reduction goal). Of course, a larger target population requires additional funding for the reentry initiative, which may be difficult to obtain.
Finally, our discussion of the different offender types released from prison everyday —such as sex offenders, drug offenders, repeat offenders, and mentally ill offenders —emphasized the need to design each phase of the reentry process to address the reintegration issues raised by the specific target population selected for reentry. Since targeting decisions will vary from jurisdiction to jurisdiction, we anticipate that reentry program models will vary from site to site as well. However, we would recommend that program developers carefully consider whether their reentry program model can address the needs of the multiple-problem offender, since it is likely that—regardless of offense-specific, offender-specific, and location-specific targeting decisions—the majority of prisoners included in their program could be described as multiple-problem offenders.