One of the hypotheses was confirmed, one was not. Subjects classified as low severity (i.e., had no diagnosable drug/alcohol problem or a drug/alcohol abuse problem) benefited equally from outpatient and residential aftercare. However, contrary to what was expected, subjects classified as high severity (i.e., were drug/alcohol dependent) also benefited equally from outpatient and residential aftercare.
The results of the chi-square analyses (Table ) showed no significant differences in the 12-month RTP rates among parolees who attended only outpatient treatment and those who attended only residential treatment, regardless of the severity of their drug/alcohol problem. The logistic regression analyses further showed that, after controlling for static demographic and criminal background variables, drug use behaviors, time spent in prison-based treatment, and time spent in aftercare (episode 1 and post-episode 1), the type of aftercare that subjects participated in (i.e., only outpatient versus only residential treatment) was not a significant predictor of 12-month RTP rates.
In both regression analyses, time spent in treatment (prison treatment and aftercare) emerged as a significant predictor of 12-month RTP. These results are consistent with previous research that has highlighted the importance of participation and retention in aftercare in combination with prison-based treatment as a means of ensuring successful treatment outcomes as measured by RTP [8
Most importantly, as it relates to the hypotheses of this study, the results of the analyses yielded no evidence of differential effectiveness between outpatient and residential aftercare in reducing recidivism among drug-involved offenders following their release from prison-based treatment, regardless of level of drug/alcohol problem severity. Within a 95% confidence interval, the odds of being returned to prison within 12 months was not appreciably different from 1.00 for those who participated in only outpatient aftercare compared to those who participated in only residential aftercare. However, other factors that characterize the treatment initiative in California, and that were not or could not be accounted or controlled for in this study, need to be considered before embracing the conclusion that there is no differential effectiveness between residential and outpatient aftercare treatment.
Most research that has demonstrated the effectiveness of prison-based treatment followed by aftercare has focused on situations where there were only one or very few in-prison treatment programs or providers and only one or very few aftercare treatment programs [9
]. Such scenarios are likely to be characterized by a high level of coordination between the in-prison and aftercare treatment providers. In addition, parolees entering aftercare are more likely to experience continuity of treatment, with aftercare treatment services picking up where prison-based treatment stopped, and are more likely to go through the in-prison and post-prison treatment experience in cohorts. Also, due to the limited number of aftercare programs/providers, the variability in the quality of the treatment services received by parolees who attend aftercare can be attenuated.
Such scenarios contrast to the situation in California, where the rapid expansion of prison-based TC treatment since 1997 has resulted in a similar rapid growth in the number of community-based programs providing treatment services to individuals paroling from the prison-based TC programs. In this study, 455 different community-based treatment programs (164 outpatient and 291 residential) delivered treatment services to the 4,165 parolees included in the analyses, who paroled from just 19 different prison-based TC programs. This computes to an average of 9.2 parolees per community-based program.
The "continuum of care" construct rests on the notion that the transition of parolees from prison-based treatment to community-based treatment be "seamless" (i.e., uninterrupted) [37
]. However, anecdotal data collected as part of the process evaluations conducted on California's prison-based treatment initiative (1997–2004) suggests that parolees entering community-based treatment programs often felt as though they were not given credit for the "uninterrupted" progress that they made in prison-based treatment, that treatment in the community did not pick up where treatment in prison left off. To the extent that this occurred, it may have contributed to increased client dissatisfaction with aftercare treatment and increased dropout rates, triggering a perception of failure on behalf of the parolee (who voluntarily entered treatment) and possibly leading to a return to criminal activity, drug use, and ultimately reincarceration.
The large number of community-based treatment programs also raises the question of variability in the quality of aftercare treatment services provided to parolees by community-based providers – a measure that is difficult to capture. In California, residential programs are required to be licensed, while non-residential (i.e., outpatient) treatment programs are not. However, neither residential nor non-residential programs are required to be certified by the state [38
], which would ensure that a program is delivering a minimal level of service quality. While CDCR and the four regionally-based SASCAs do take into consideration the quality of treatment services provided by individual community-based treatment programs when making referrals and placement decisions, the current system for transitioning parolees from prison-based treatment programs to the large network of community-based treatment programs was not designed to assess the quality of aftercare treatment services, and it does not ensure that community-based treatment programs take into account progress that parolees made in prison-based treatment.
Finally, the process of transitioning parolees from prison- to community-based treatment in California also does not include a formal systematic assessment process for matching parolees' needs with community-based treatment programs or services. Subjects in this study consisted of parolees who attended only residential or only outpatient aftercare. The majority of these subjects (68%) attended only residential aftercare. There were an additional 1,960 parolees who attended a combination of both residential and outpatient aftercare following their release from prison and who were not included in this study. Residential treatment constituted the first aftercare treatment episode for 92% of these 1,960 parolees. Combined, these facts indicate that, following conventional wisdom, parolees were most often referred to and encouraged to attend residential treatment in the community (i.e., the more intensive treatment modality).
There currently exist no validated assessments that are designed specifically for substance-abusing parolees who are encouraged or required to participate in treatment after they are released from prison. However, recently, as part of the ongoing NIDA-funded Criminal Justice – Drug Abuse Treatment Studies (CJ-DATS) research initiative, researchers at several sites across the country are testing the efficacy of the Inmate Pre-Release Assessment
] to match paroling offenders to an appropriate modality of aftercare. Specifically designed as a pre-release risk measure for prison-based substance abuse treatment graduates, the IPASS takes into account inmates' historical drug use and criminal activity, as well as performance in prison-based treatment. If successful, the IPASS instrument will constitute the first validated tool for assessing treatment needs of substance abusing parolees being released from prison and for guiding referrals to effective aftercare treatment.
With respect to the limitations of this study, the measure of alcohol/drug problem severity, as well as some measures relating to demographic background and criminal background, were based on self-report. Although questions are often raised about self-report data, prior research indicates that self-report interviews, when properly conducted, are generally reliable and valid in measuring drug and alcohol use [40
] and criminal involvement [43
]. In addition, some self-report measures were compared to official records. Where this occurred, analyses comparing self-report data to those obtained from the official records showed no differences (e.g., type of offense). Combined, these factors alleviated concerns regarding the veracity of self-report data.
This study was concerned with assessing the differential effectiveness of residential versus outpatient aftercare treatment among those subjects who actually participated in aftercare treatment. Thus, the potential impact of selection bias on the results is limited to factors that may have influenced the decision to participate in residential or outpatient aftercare, but not the decision to participate or not participate in aftercare. Despite the emphasis on referring parolees to residential treatment, most parolees do play a role in choosing the modality of aftercare treatment that they participate in (i.e., residential or outpatient). With respect to selection bias and its impact on the choice of treatment modality, the logistic regression analyses controlled for a full array of variables that likely influence the decision regarding which modality of treatment to participate in. However, there may be other unknown individual- and system-level factors that were not measured in this study, and thus not controlled for, that influence the decision to participate in residential or outpatient aftercare treatment.
Finally, although the logistic regression analyses accounted for the clustered design and the resulting variability among participants within and between the different in-prison treatment programs, relevant program-level variables were not collected and thus were outside the scope of this study. Future studies should examine various programmatic differences using multilevel modeling techniques to explain the variation between different in-prison treatment programs (e.g., quality of the individual treatment program).