Bernadette Pelissier, Federal Bureau of Prisons
Timothy Cadigan, Administrative Office of the U.S. Courts

Introduction
Purpose of the Study
Treatment Services
Research Questions
Data
Methods
Results
Discussion

THREE QUARTERS of the inmates housed in state and federal prison in 1997 could be characterized as alcohol or drug-involved (Mumola, 1999). Yet only approximately 20 percent of those within 6 months of release report having received treatment (Mumola, 1999). Drug offenders, many of whom may be drug users, account for an increasing percentage of those released from prison. In 1980, 11 percent of all releasees were drug-involved. In 1990, they accounted for 26 percent of releasees and this percentage increased to 32 percent in 1998 (Lynch & Sabol, 2001; Travis, Solomon, & Waul, 2001).

Among those who have evaluated the effectiveness of prison-based drug treatment programs, considerable discussion has arisen about the need to provide aftercare services and ensure continuity of care. The effects of in-prison treatment might not be maintained without continuity of care after release (Field, 1998; Hiller, Knight, & Simpson, 1999; Knight, Simpson, & Hiller, 1999; Martin, Butzin, Saum, & Inciardi, 1999; Swartz, Lurigio, & Slomka, 1996; Wexler, Melnick, Lowe, & Peters, 1999). The recent focus upon reentry (Travis et al., 2001) draws attention to the issue of treatment after release from prison by suggesting that treatment contact be initiated before releasing an individual from prison. But more generally, the emphasis on reentry issues calls for collaboration between various criminal justice agencies (Burke, 2001).

Because some prison systems operate within a context where post-prison responsibilities fall to an agency other than the agency responsible for an individual during incarceration, the issues surrounding continuity of care are often times unknown. The federal criminal justice system is such an example. In the federal criminal justice system, unlike some state, county and local jurisdictions, the incarcerating agency (Bureau of Prisons (BOP)) is not only in a different agency from the supervision provider (Administrative Office of the U.S. Courts (AOUSC)), but it is also in an entirely different branch of government. The BOP is in the executive branch whereas AOUSC is in the judicial branch.

The federal criminal justice system has been increasing its focus on the reentry process in response to influences from both within and outside the federal criminal justice system. Those internal influences include the Administrative Office of the United States Courts (AOUSC), the Department of Justice (DOJ), and the Bureau of Prisons (BOP), while the external influences include Congress and the criminal justice literature that has focused on the value of a true reentry approach. Overcoming the cultural, procedural and systematic differences to develop an effective reentry infrastructure and process will require tremendous commitment and creativity from staff in both agencies.

back to top

Purpose of Study

A long-standing memorandum of understanding (MOU) initiated between AOUSC and BOP in 1992 concerns the handling of substance abuse treatment cases and more specifically the transition of those cases from the BOP to the AOUSC. That MOU recognized the importance of the reentry phase to the success of the substance treatment provided by both agencies long before that recognition was in vogue in the criminal justice literature. Essentially, the purpose of this study is to provide a measure of impact of that MOU on post-release substance abuse treatment.

Drug treatment services and continuity of care are examined among a cohort of approximately 25,000 individuals released in 1999 from the BOP to supervision by a U.S. probation officer. The tracking of treatment received during and after incarceration was facilitated by an interagency agreement allowing for data sharing between these two federal agencies. Information on drug treatment services includes prison-based residential treatment, outpatient drug treatment during halfway house placements and drug treatment provided while under post-release supervision. Models are developed to predict drug aftercare and the analyses include an assessment of the extent to which aftercare is prioritized for those who completed residential drug treatment while incarcerated. Recommendations for interagency information sharing and service planning will be discussed.

back to top

Treatment Services

The BOP's residential drug treatment program (RDAP) provides 500 hours of treatment over a 9-month time frame. The RDAP participants must meet established admission criteria, which include a documented substance abuse problem and a willingness to partake fully in treatment services. Individuals admitted to RDAP must usually be within 36 months of release. The program uses a cognitive-behavioral treatment model and treatment is generally provided for a half-day, five days per week. The program attempts to identify, confront, and alter the attitudes and thinking patterns that led to criminal behaviors and drug use.

Transitional services (TS) is the other major component of the BOP's drug treatment program. This focuses on providing outpatient treatment during a halfway house placement, for those who receive such a placement. Both those who participated in the residential prison-based treatment and those who did not can make use of Transitional Services. Thus, prior participation in RDAP is not a requirement for receiving TS treatment.

Once received by the AOUSC for supervised release or parole supervision, offenders can be required by judicial order to participate in the substance abuse treatment program (SATP). The AOUSC SATP involves many components, including assessment, outpatient treatment, inpatient treatment, detoxification, methadone maintenance, and a variety of methodologies to detect substance use. Within the components of the SATP, which is primarily an abstinence program, the actual treatment provided, whether inpatient or outpatient, employs a wide range of modalities from cognitive to behavior modification to self help and others.

back to top

Research Questions

For purposes of assessing continuity of care upon release to supervision by a probation officer, we selected those who received at least one of the two primary methods of substance abuse treatment provided to inmates in the BOP: Residential Drug Abuse Program (RDAP) and Transitional Services (TS).

An assessment of continuity of care first requires information on who receives treatment during incarceration and an understanding of the numbers who participated in one or both of the BOP programs and who completed or failed to complete the program. Individuals may have received treatment during their incarceration in prison but not during a halfway house placement (some because they did not receive a halfway house placement). On the other hand, individuals may first begin receiving treatment during their halfway house placement. Therefore, the first question to be examined is: What percent of those among a release cohort from BOP custody received RDAP and/or TS services?

The first level of assessment of continuity of care is to identify the percentage of those receiving treatment during BOP custody who also received treatment post-release. A primary interest is in determining whether, for example, treatment completers are more likely to receive treatment under supervision. However, other questions of interest include identifying subgroups that are more or less likely to receive post-release treatment. For example, are women more or less likely than men to receive post-release treatment? What evidence is there for selecting individuals for post-release treatment based on their background characteristics?

Anecdotal information suggests that the philosophies of judges concerning drug treatment can vary considerably. Therefore, another question of great interest is whether there are district or circuit differences in the provision of aftercare services.

back to top

Data

One particularly vexing problem has been obtaining information on continuity of care from the onset of substance abuse treatment an offender receives in the BOP to the substance abuse treatment that offender receives once released to the supervision of a United States probation officer. The BOP maintains an operational database, known as Sentry, where substance abuse treatment records are stored. Likewise the AOUSC maintains an operational case management database where substance abuse treatment records are stored, known as the National Treatment Database (NTD). To study the continuity of substance abuse treatment care in the federal criminal justice system required the merging of data from those systems. While this is a relatively simple and straightforward concept, the reality proved to be more complicated.

Those complications were ultimately overcome and we were able to take Sentry records for 27,420 individuals released from BOP custody in 1999 and match most of them to NTD supervised release/parole supervision records of the AOUSC. The data was successfully merged for 26,813 cases for a matching rate of 98 percent which was deemed satisfactory for purposes of this research. The majority of cases were matched by FBI number (89.6 percent). The remaining cases were matched by Register Number (6.8 percent), social security number (3.0 percent), date of birth, last name (1st 4 letters) and first name (1st 3 letters) (0.6 percent). Sex and race could not be used for matching purposes because of known differences in coding.

Next we identified background characteristics for which we wanted to assess whether there was an association with higher or lower levels of post-release treatment. The factors obtained from the BOP's Sentry data base included: gender, race, ethnicity, age at time of release, type of drug treatment received, and items indicating whether the sentence was served only in a halfway house and whether the RDAP and/or TS services were received in a previous incarceration. Two criminal history indicators were obtained from the AOUSC database: history of felony and risk prediction index (RPI) score. The RPI is a risk prediction index that uses information about an offender, including prior criminal record, to estimate the likelihood that the offender will recidivate during his or her term of supervision.

Since we were also interested in assessing district differences, we obtained information on the caseload size for each district from the AOUSC database. The federal judicial system is comprised of 94 federal districts with 93 operational probation offices. Probation cases from the 94th district, Northern Mariana Islands, are handled by the probation office within the District of Guam. Those 94 districts are grouped into 12 regional judicial circuits. Anecdotal information over the years has suggested that substance abuse cases can be handled differently both from circuit to circuit and from district to district. We tested those assumptions with this dataset.

back to top

Methods

Hierarchical linear modeling (HLM) was used to analyze the data (Golstein, 1995; Raudenbush, Bryk, Cheong, & Congdon, 2000). This statistical technique allows simultaneous modeling of individual level effects (e.g., gender, race, risk score) as well as group level effects (e.g., district and circuit). This statistical technique allows us to answer the question of whether there are district and circuit differences in the provision of treatment after controlling for individual background characteristics. In addition to the individual characteristics identified in the data section, we also classified individuals into eight categories of in-prison/halfway house treatment and used these categories as predictors in our analytic models. These categories represent the various combinations of treatment failure or completion for either or both RDAP and TS (see Table 1 below for a listing of the categories).

back to top

Results

We begin by identifying those among the 26,813 individuals released to supervision in 1999 who received RDAP or TS treatment while under BOP custody. Approximately 20 percent (n=5320) of those released to supervision received either in-prison residential drug treatment (RDAP) or transitional services (TS) during a halfway house placement. All of the analyses we report are limited to these 5320 individuals. Table 1 provides information by completion status on the numbers of those receiving one or both of these services. More than half of those who received drug treatment completed both the in-prison and the halfway house treatment components. Another 14 percent completed RDAP but either they did not receive transitional services (5.9 percent) or they did receive TS but did not successfully complete it (8.4 percent). We note that many of those who completed RDAP but did not receive TS did not receive a halfway house placement and thus could not receive TS. The majority of those who failed RDAP also did not receive TS. While 10.2 percent of those receiving treatment were RDAP failures who did not receive TS, only 1 percent were RDAP failures who received such services. Finally, more than 18 percent of those receiving one or both types of services received services only during their halfway house placement (e.g., TS). We note that an additional 1 percent of the release cohort (n=294) received RDAP and/or TS in a previous incarceration. Many of these individuals were revoked and served an insufficient amount of time to be readmitted to RDAP.

Having developed the profile of the 5,320 offenders treated by the BOP, we now turn to the reentry aspect of continuity of care for those offenders as they are released from BOP to AOUSC supervision. Table 2 provides information on treatment services received while under supervision for all those individuals who received treatment under BOP custody. Information is provided for each of the BOP treatment categories contained in Table 1.

Because treatment can be initiated at any point during supervision, we defined continuity of care as treatment assigned by a probation officer within 90 days after admission to supervision. This ninety-day period was selected for a variety of reasons. First, we felt that the continuum of care would be broken beyond ninety days. Second, the majority of individuals who suddenly begin treatment more than ninety days after release from prison were likely to do so in response to substance use or other violation behavior, which seemed to place them beyond the purposes of this project. In addition, detoxification services were also likely to be in response to substance abuse, and so were also excluded, even if initiated less than 90 days after admission to supervision.

Overall, 37.6 percent of those who received treatment under BOP custody also received treatment within 90 days after admission to supervision. An additional 15 percent of these individuals received detoxification services or inpatient/outpatient services which started more than 90 days after admission to supervision.

Table 2 shows that for six of the eight BOP treatment types, the percentage who received treatment during supervision ranged between 39 and 43 percent. The two categories that stood out showed lower percentages receiving treatment while under supervision: 33.6 percent among those who failed RDAP and did not receive TS and 16.1 percent among those who completed RDAP and did not receive TS. It appears that both groups who did not receive TS had lower rates of continuity of care from BOP to AOUSC.

We next assessed the role of individual characteristics, including background factors and type of BOP treatment received, as well as the role of group level characteristics (district and circuit) using HML multivariate analyses. Table 3 summarizes the findings by indicating whether a predictor was significant and if so, whether the relationship was positive or negative. Predictors that were positive are those associated with a greater probability of receiving treatment during supervised release and predictors that were negative are those associated with a lower probability of receiving such treatment.

The offender's race showed no significant impact as a predictor of receiving post-release treatment. Prior felony and risk prediction index (RPI) scores were shown to have a positive predictive value, which demonstrates that controlling for other factors, offenders are more likely to receive treatment post-release when they have prior felonies or present a greater risk to the community. Age at release had a negative impact on the likelihood of treatment post-release, a finding consistent with anticipated or desired outcomes for the SATP.

Two demographic factors did have somewhat unanticipated predictive power: gender and ethnicity. All other factors being equal, female offenders are more likely to received post-release substance abuse treatment. This is an interesting although somewhat unexpected finding which may require further study and explanation. The one somewhat surprising finding among the demographic characteristics is ethnicity. Hispanic ethnicity, controlling for race, has a negative impact on the likelihood of post-release treatment. This would appear to be an area requiring further study and analysis to determine causal factors so that if appropriate these factors can be addressed.

Only one category of BOP treatment received was associated with a lower likelihood of SATP treatment than on average: RDAP Complete – No TS. In contrast, those who completed RDAP and received TS, whether or not they completed TS, were more likely to receive SATP treatment. In addition, those who received TS only and completed were more likely to receive SATP treatment. These results indicate the need to identify why those individuals, albeit a small percentage of those receiving BOP treatment, who received in-prison treatment only, have a significantly lower likelihood of receiving post-release treatment.

The last set of results indicates that our only predictor related to district characteristics, caseload size, was not significant. However, it is important to note that HLM results are valuable in that they provide information on unexplained variation. After controlling for the individual characteristics, the HLM results indicate that there is variation left to explain at both the district and circuit level. This implies that individuals with similar characteristics are more likely to receive treatment in some districts or circuits but less likely to receive treatment in other districts or circuits.

back to top

Discussion

The goal of this research was to assess continuity of care across two independent substance abuse treatment programs, each with its own goals and purposes, and identify potential policy changes that would enhance the combined impact of these programs on the offenders they serve. This approach represents a major philosophical shift from the past operation of these two programs. However, the agencies have a long-recognized understanding of their interrelationship, particularly in the area of substance abuse treatment, which should sustain them on the long journey ahead. That journey requires that all staff, from senior managers to frontline officers and case managers, recognize that our ultimate success is contingent on a systemic approach to the problems posed by substance abusing offenders. Our research indicates that changing certain policies would likely improve the federal "system" of substance abuse treatment for offenders. Future research will be required to examine the impact of these policies.

It should be noted that while the research did identify issues that should be addressed, the majority of factors were confirmed in the direction that policy makers and practitioners would want. Specifically, race was shown to have no impact on whether or not someone receives substance abuse treatment. Prior felony convictions and RPI score (both predictors of the offender's risk to the community) were shown to have a positive relationship with the likelihood of treatment. Age at time of release showed a negative impact. All of these outcomes bode well for the AOUSC SATP and support that it is a strong program which consistently directs its resources toward the offenders of most concern in a public safety program. However, like any good program it can be improved and the following steps by both agencies would enhance the program.

The AOUSC needs to identify a policy that could decrease variation across districts and circuits in the provision of drug treatment to those who received treatment during incarceration by the BOP. The models employed in this research showed significant variation by district and circuit even after controlling for caseload size and related variables. To insure a consistent national substance abuse program for offenders, regardless of their district or circuit of release, the treatment they receive must have less unexplainable variation.

The identification and implementation by the BOP of policies to increase the likelihood that RDAP offenders receive TS treatment would have several advantages. This would not only maintain the continuity of care but would also increase the likelihood that, once released, offenders would receive treatment during supervision.

back to top